diff options
Diffstat (limited to 'arch/powerpc/platforms/pseries')
52 files changed, 24365 insertions, 0 deletions
diff --git a/arch/powerpc/platforms/pseries/Kconfig b/arch/powerpc/platforms/pseries/Kconfig new file mode 100644 index 000000000..4ebf2ef28 --- /dev/null +++ b/arch/powerpc/platforms/pseries/Kconfig @@ -0,0 +1,186 @@ +# SPDX-License-Identifier: GPL-2.0 +config PPC_PSERIES + depends on PPC64 && PPC_BOOK3S + bool "IBM pSeries & new (POWER5-based) iSeries" + select HAVE_PCSPKR_PLATFORM + select MPIC + select OF_DYNAMIC + select FORCE_PCI + select PCI_MSI + select GENERIC_ALLOCATOR + select PPC_XICS + select PPC_XIVE_SPAPR + select PPC_ICP_NATIVE + select PPC_ICP_HV + select PPC_ICS_RTAS + select PPC_I8259 + select PPC_RTAS + select PPC_RTAS_DAEMON + select RTAS_ERROR_LOGGING + select PPC_UDBG_16550 + select PPC_DOORBELL + select HOTPLUG_CPU + select FORCE_SMP + select SWIOTLB + select ARCH_SUPPORTS_PER_VMA_LOCK + default y + +config PARAVIRT + bool + +config PARAVIRT_SPINLOCKS + bool + +config PARAVIRT_TIME_ACCOUNTING + select PARAVIRT + bool + +config PPC_SPLPAR + bool "Support for shared-processor logical partitions" + depends on PPC_PSERIES + select PARAVIRT_SPINLOCKS if PPC_QUEUED_SPINLOCKS + select PARAVIRT_TIME_ACCOUNTING if VIRT_CPU_ACCOUNTING_GEN + default y + help + Enabling this option will make the kernel run more efficiently + on logically-partitioned pSeries systems which use shared + processors, that is, which share physical processors between + two or more partitions. + + Say Y if you are unsure. + +config DTL + bool "Dispatch Trace Log" + depends on PPC_SPLPAR && DEBUG_FS + help + SPLPAR machines can log hypervisor preempt & dispatch events to a + kernel buffer. Saying Y here will enable logging these events, + which are accessible through a debugfs file. + + Say N if you are unsure. + +config PSERIES_ENERGY + tristate "pSeries energy management capabilities driver" + depends on PPC_PSERIES + default y + help + Provides interface to platform energy management capabilities + on supported PSERIES platforms. + Provides: /sys/devices/system/cpu/pseries_(de)activation_hint_list + and /sys/devices/system/cpu/cpuN/pseries_(de)activation_hint + +config IO_EVENT_IRQ + bool "IO Event Interrupt support" + depends on PPC_PSERIES + default y + help + Select this option, if you want to enable support for IO Event + interrupts. IO event interrupt is a mechanism provided by RTAS + to return information about hardware error and non-error events + which may need OS attention. RTAS returns events for multiple + event types and scopes. Device drivers can register their handlers + to receive events. + + This option will only enable the IO event platform code. You + will still need to enable or compile the actual drivers + that use this infrastructure to handle IO event interrupts. + + Say Y if you are unsure. + +config LPARCFG + bool "LPAR Configuration Data" + depends on PPC_PSERIES + help + Provide system capacity information via human readable + <key word>=<value> pairs through a /proc/ppc64/lparcfg interface. + +config PPC_PSERIES_DEBUG + depends on PPC_PSERIES && PPC_EARLY_DEBUG + bool "Enable extra debug logging in platforms/pseries" + default y + help + Say Y here if you want the pseries core to produce a bunch of + debug messages to the system log. Select this if you are having a + problem with the pseries core and want to see more of what is + going on. This does not enable debugging in lpar.c, which must + be manually done due to its verbosity. + +config PPC_SMLPAR + bool "Support for shared-memory logical partitions" + depends on PPC_PSERIES + select LPARCFG + help + Select this option to enable shared memory partition support. + With this option a system running in an LPAR can be given more + memory than physically available and will allow firmware to + balance memory across many LPARs. + +config CMM + tristate "Collaborative memory management" + depends on PPC_SMLPAR + select MEMORY_BALLOON + default y + help + Select this option, if you want to enable the kernel interface + to reduce the memory size of the system. This is accomplished + by allocating pages of memory and put them "on hold". This only + makes sense for a system running in an LPAR where the unused pages + will be reused for other LPARs. The interface allows firmware to + balance memory across many LPARs. + +config HV_PERF_CTRS + bool "Hypervisor supplied PMU events (24x7 & GPCI)" + default y + depends on PERF_EVENTS && PPC_PSERIES + help + Enable access to hypervisor supplied counters in perf. Currently, + this enables code that uses the hcall GetPerfCounterInfo and 24x7 + interfaces to retrieve counters. GPCI exists on Power 6 and later + systems. 24x7 is available on Power 8 and later systems. + + If unsure, select Y. + +config IBMVIO + depends on PPC_PSERIES + bool + default y + +config IBMEBUS + depends on PPC_PSERIES && !CPU_LITTLE_ENDIAN + bool "Support for GX bus based adapters" + help + Bus device driver for GX bus based adapters. + +config PSERIES_PLPKS + depends on PPC_PSERIES + select NLS + bool + # PowerVM provides an isolated Platform Keystore (PKS) storage + # allocation for each LPAR with individually managed access + # controls to store sensitive information securely. It can be + # used to store asymmetric public keys or secrets as required + # by different usecases. + # + # This option is selected by in-kernel consumers that require + # access to the PKS. + +config PAPR_SCM + depends on PPC_PSERIES && MEMORY_HOTPLUG && LIBNVDIMM + tristate "Support for the PAPR Storage Class Memory interface" + help + Enable access to hypervisor provided storage class memory. + +config PPC_SVM + bool "Secure virtual machine (SVM) support for POWER" + depends on PPC_PSERIES + select SWIOTLB + select ARCH_HAS_MEM_ENCRYPT + select ARCH_HAS_FORCE_DMA_UNENCRYPTED + select ARCH_HAS_CC_PLATFORM + help + There are certain POWER platforms which support secure guests using + the Protected Execution Facility, with the help of an Ultravisor + executing below the hypervisor layer. This enables support for + those guests. + + If unsure, say "N". diff --git a/arch/powerpc/platforms/pseries/Makefile b/arch/powerpc/platforms/pseries/Makefile new file mode 100644 index 000000000..53c3b91af --- /dev/null +++ b/arch/powerpc/platforms/pseries/Makefile @@ -0,0 +1,39 @@ +# SPDX-License-Identifier: GPL-2.0 +ccflags-$(CONFIG_PPC64) := $(NO_MINIMAL_TOC) +ccflags-$(CONFIG_PPC_PSERIES_DEBUG) += -DDEBUG + +obj-y := lpar.o hvCall.o nvram.o reconfig.o \ + of_helpers.o rtas-work-area.o papr-sysparm.o \ + setup.o iommu.o event_sources.o ras.o \ + firmware.o power.o dlpar.o mobility.o rng.o \ + pci.o pci_dlpar.o eeh_pseries.o msi.o \ + papr_platform_attributes.o dtl.o +obj-$(CONFIG_SMP) += smp.o +obj-$(CONFIG_KEXEC_CORE) += kexec.o +obj-$(CONFIG_PSERIES_ENERGY) += pseries_energy.o + +obj-$(CONFIG_HOTPLUG_CPU) += hotplug-cpu.o +obj-$(CONFIG_MEMORY_HOTPLUG) += hotplug-memory.o pmem.o + +obj-$(CONFIG_HVC_CONSOLE) += hvconsole.o +obj-$(CONFIG_HVCS) += hvcserver.o +obj-$(CONFIG_HCALL_STATS) += hvCall_inst.o +obj-$(CONFIG_CMM) += cmm.o +obj-$(CONFIG_IO_EVENT_IRQ) += io_event_irq.o +obj-$(CONFIG_LPARCFG) += lparcfg.o +obj-$(CONFIG_IBMVIO) += vio.o +obj-$(CONFIG_IBMEBUS) += ibmebus.o +obj-$(CONFIG_PAPR_SCM) += papr_scm.o +obj-$(CONFIG_PPC_SPLPAR) += vphn.o +obj-$(CONFIG_PPC_SVM) += svm.o +obj-$(CONFIG_FA_DUMP) += rtas-fadump.o +obj-$(CONFIG_PSERIES_PLPKS) += plpks.o +obj-$(CONFIG_PPC_SECURE_BOOT) += plpks-secvar.o +obj-$(CONFIG_SUSPEND) += suspend.o +obj-$(CONFIG_PPC_VAS) += vas.o vas-sysfs.o + +obj-$(CONFIG_ARCH_HAS_CC_PLATFORM) += cc_platform.o + +# nothing that operates in real mode is safe for KASAN +KASAN_SANITIZE_ras.o := n +KASAN_SANITIZE_kexec.o := n diff --git a/arch/powerpc/platforms/pseries/cc_platform.c b/arch/powerpc/platforms/pseries/cc_platform.c new file mode 100644 index 000000000..e8021af83 --- /dev/null +++ b/arch/powerpc/platforms/pseries/cc_platform.c @@ -0,0 +1,26 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Confidential Computing Platform Capability checks + * + * Copyright (C) 2021 Advanced Micro Devices, Inc. + * + * Author: Tom Lendacky <thomas.lendacky@amd.com> + */ + +#include <linux/export.h> +#include <linux/cc_platform.h> + +#include <asm/machdep.h> +#include <asm/svm.h> + +bool cc_platform_has(enum cc_attr attr) +{ + switch (attr) { + case CC_ATTR_MEM_ENCRYPT: + return is_secure_guest(); + + default: + return false; + } +} +EXPORT_SYMBOL_GPL(cc_platform_has); diff --git a/arch/powerpc/platforms/pseries/cmm.c b/arch/powerpc/platforms/pseries/cmm.c new file mode 100644 index 000000000..5f4037c1d --- /dev/null +++ b/arch/powerpc/platforms/pseries/cmm.c @@ -0,0 +1,663 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Collaborative memory management interface. + * + * Copyright (C) 2008 IBM Corporation + * Author(s): Brian King (brking@linux.vnet.ibm.com), + */ + +#include <linux/ctype.h> +#include <linux/delay.h> +#include <linux/errno.h> +#include <linux/fs.h> +#include <linux/gfp.h> +#include <linux/kthread.h> +#include <linux/module.h> +#include <linux/oom.h> +#include <linux/reboot.h> +#include <linux/sched.h> +#include <linux/stringify.h> +#include <linux/swap.h> +#include <linux/device.h> +#include <linux/balloon_compaction.h> +#include <asm/firmware.h> +#include <asm/hvcall.h> +#include <asm/mmu.h> +#include <linux/uaccess.h> +#include <linux/memory.h> +#include <asm/plpar_wrappers.h> + +#include "pseries.h" + +#define CMM_DRIVER_VERSION "1.0.0" +#define CMM_DEFAULT_DELAY 1 +#define CMM_HOTPLUG_DELAY 5 +#define CMM_DEBUG 0 +#define CMM_DISABLE 0 +#define CMM_OOM_KB 1024 +#define CMM_MIN_MEM_MB 256 +#define KB2PAGES(_p) ((_p)>>(PAGE_SHIFT-10)) +#define PAGES2KB(_p) ((_p)<<(PAGE_SHIFT-10)) + +#define CMM_MEM_HOTPLUG_PRI 1 + +static unsigned int delay = CMM_DEFAULT_DELAY; +static unsigned int hotplug_delay = CMM_HOTPLUG_DELAY; +static unsigned int oom_kb = CMM_OOM_KB; +static unsigned int cmm_debug = CMM_DEBUG; +static unsigned int cmm_disabled = CMM_DISABLE; +static unsigned long min_mem_mb = CMM_MIN_MEM_MB; +static bool __read_mostly simulate; +static unsigned long simulate_loan_target_kb; +static struct device cmm_dev; + +MODULE_AUTHOR("Brian King <brking@linux.vnet.ibm.com>"); +MODULE_DESCRIPTION("IBM System p Collaborative Memory Manager"); +MODULE_LICENSE("GPL"); +MODULE_VERSION(CMM_DRIVER_VERSION); + +module_param_named(delay, delay, uint, 0644); +MODULE_PARM_DESC(delay, "Delay (in seconds) between polls to query hypervisor paging requests. " + "[Default=" __stringify(CMM_DEFAULT_DELAY) "]"); +module_param_named(hotplug_delay, hotplug_delay, uint, 0644); +MODULE_PARM_DESC(hotplug_delay, "Delay (in seconds) after memory hotplug remove " + "before loaning resumes. " + "[Default=" __stringify(CMM_HOTPLUG_DELAY) "]"); +module_param_named(oom_kb, oom_kb, uint, 0644); +MODULE_PARM_DESC(oom_kb, "Amount of memory in kb to free on OOM. " + "[Default=" __stringify(CMM_OOM_KB) "]"); +module_param_named(min_mem_mb, min_mem_mb, ulong, 0644); +MODULE_PARM_DESC(min_mem_mb, "Minimum amount of memory (in MB) to not balloon. " + "[Default=" __stringify(CMM_MIN_MEM_MB) "]"); +module_param_named(debug, cmm_debug, uint, 0644); +MODULE_PARM_DESC(debug, "Enable module debugging logging. Set to 1 to enable. " + "[Default=" __stringify(CMM_DEBUG) "]"); +module_param_named(simulate, simulate, bool, 0444); +MODULE_PARM_DESC(simulate, "Enable simulation mode (no communication with hw)."); + +#define cmm_dbg(...) if (cmm_debug) { printk(KERN_INFO "cmm: "__VA_ARGS__); } + +static atomic_long_t loaned_pages; +static unsigned long loaned_pages_target; +static unsigned long oom_freed_pages; + +static DEFINE_MUTEX(hotplug_mutex); +static int hotplug_occurred; /* protected by the hotplug mutex */ + +static struct task_struct *cmm_thread_ptr; +static struct balloon_dev_info b_dev_info; + +static long plpar_page_set_loaned(struct page *page) +{ + const unsigned long vpa = page_to_phys(page); + unsigned long cmo_page_sz = cmo_get_page_size(); + long rc = 0; + int i; + + if (unlikely(simulate)) + return 0; + + for (i = 0; !rc && i < PAGE_SIZE; i += cmo_page_sz) + rc = plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_LOANED, vpa + i, 0); + + for (i -= cmo_page_sz; rc && i != 0; i -= cmo_page_sz) + plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_ACTIVE, + vpa + i - cmo_page_sz, 0); + + return rc; +} + +static long plpar_page_set_active(struct page *page) +{ + const unsigned long vpa = page_to_phys(page); + unsigned long cmo_page_sz = cmo_get_page_size(); + long rc = 0; + int i; + + if (unlikely(simulate)) + return 0; + + for (i = 0; !rc && i < PAGE_SIZE; i += cmo_page_sz) + rc = plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_ACTIVE, vpa + i, 0); + + for (i -= cmo_page_sz; rc && i != 0; i -= cmo_page_sz) + plpar_hcall_norets(H_PAGE_INIT, H_PAGE_SET_LOANED, + vpa + i - cmo_page_sz, 0); + + return rc; +} + +/** + * cmm_alloc_pages - Allocate pages and mark them as loaned + * @nr: number of pages to allocate + * + * Return value: + * number of pages requested to be allocated which were not + **/ +static long cmm_alloc_pages(long nr) +{ + struct page *page; + long rc; + + cmm_dbg("Begin request for %ld pages\n", nr); + + while (nr) { + /* Exit if a hotplug operation is in progress or occurred */ + if (mutex_trylock(&hotplug_mutex)) { + if (hotplug_occurred) { + mutex_unlock(&hotplug_mutex); + break; + } + mutex_unlock(&hotplug_mutex); + } else { + break; + } + + page = balloon_page_alloc(); + if (!page) + break; + rc = plpar_page_set_loaned(page); + if (rc) { + pr_err("%s: Can not set page to loaned. rc=%ld\n", __func__, rc); + __free_page(page); + break; + } + + balloon_page_enqueue(&b_dev_info, page); + atomic_long_inc(&loaned_pages); + adjust_managed_page_count(page, -1); + nr--; + } + + cmm_dbg("End request with %ld pages unfulfilled\n", nr); + return nr; +} + +/** + * cmm_free_pages - Free pages and mark them as active + * @nr: number of pages to free + * + * Return value: + * number of pages requested to be freed which were not + **/ +static long cmm_free_pages(long nr) +{ + struct page *page; + + cmm_dbg("Begin free of %ld pages.\n", nr); + while (nr) { + page = balloon_page_dequeue(&b_dev_info); + if (!page) + break; + plpar_page_set_active(page); + adjust_managed_page_count(page, 1); + __free_page(page); + atomic_long_dec(&loaned_pages); + nr--; + } + cmm_dbg("End request with %ld pages unfulfilled\n", nr); + return nr; +} + +/** + * cmm_oom_notify - OOM notifier + * @self: notifier block struct + * @dummy: not used + * @parm: returned - number of pages freed + * + * Return value: + * NOTIFY_OK + **/ +static int cmm_oom_notify(struct notifier_block *self, + unsigned long dummy, void *parm) +{ + unsigned long *freed = parm; + long nr = KB2PAGES(oom_kb); + + cmm_dbg("OOM processing started\n"); + nr = cmm_free_pages(nr); + loaned_pages_target = atomic_long_read(&loaned_pages); + *freed += KB2PAGES(oom_kb) - nr; + oom_freed_pages += KB2PAGES(oom_kb) - nr; + cmm_dbg("OOM processing complete\n"); + return NOTIFY_OK; +} + +/** + * cmm_get_mpp - Read memory performance parameters + * + * Makes hcall to query the current page loan request from the hypervisor. + * + * Return value: + * nothing + **/ +static void cmm_get_mpp(void) +{ + const long __loaned_pages = atomic_long_read(&loaned_pages); + const long total_pages = totalram_pages() + __loaned_pages; + int rc; + struct hvcall_mpp_data mpp_data; + signed long active_pages_target, page_loan_request, target; + signed long min_mem_pages = (min_mem_mb * 1024 * 1024) / PAGE_SIZE; + + if (likely(!simulate)) { + rc = h_get_mpp(&mpp_data); + if (rc != H_SUCCESS) + return; + page_loan_request = div_s64((s64)mpp_data.loan_request, + PAGE_SIZE); + target = page_loan_request + __loaned_pages; + } else { + target = KB2PAGES(simulate_loan_target_kb); + page_loan_request = target - __loaned_pages; + } + + if (target < 0 || total_pages < min_mem_pages) + target = 0; + + if (target > oom_freed_pages) + target -= oom_freed_pages; + else + target = 0; + + active_pages_target = total_pages - target; + + if (min_mem_pages > active_pages_target) + target = total_pages - min_mem_pages; + + if (target < 0) + target = 0; + + loaned_pages_target = target; + + cmm_dbg("delta = %ld, loaned = %lu, target = %lu, oom = %lu, totalram = %lu\n", + page_loan_request, __loaned_pages, loaned_pages_target, + oom_freed_pages, totalram_pages()); +} + +static struct notifier_block cmm_oom_nb = { + .notifier_call = cmm_oom_notify +}; + +/** + * cmm_thread - CMM task thread + * @dummy: not used + * + * Return value: + * 0 + **/ +static int cmm_thread(void *dummy) +{ + unsigned long timeleft; + long __loaned_pages; + + while (1) { + timeleft = msleep_interruptible(delay * 1000); + + if (kthread_should_stop() || timeleft) + break; + + if (mutex_trylock(&hotplug_mutex)) { + if (hotplug_occurred) { + hotplug_occurred = 0; + mutex_unlock(&hotplug_mutex); + cmm_dbg("Hotplug operation has occurred, " + "loaning activity suspended " + "for %d seconds.\n", + hotplug_delay); + timeleft = msleep_interruptible(hotplug_delay * + 1000); + if (kthread_should_stop() || timeleft) + break; + continue; + } + mutex_unlock(&hotplug_mutex); + } else { + cmm_dbg("Hotplug operation in progress, activity " + "suspended\n"); + continue; + } + + cmm_get_mpp(); + + __loaned_pages = atomic_long_read(&loaned_pages); + if (loaned_pages_target > __loaned_pages) { + if (cmm_alloc_pages(loaned_pages_target - __loaned_pages)) + loaned_pages_target = __loaned_pages; + } else if (loaned_pages_target < __loaned_pages) + cmm_free_pages(__loaned_pages - loaned_pages_target); + } + return 0; +} + +#define CMM_SHOW(name, format, args...) \ + static ssize_t show_##name(struct device *dev, \ + struct device_attribute *attr, \ + char *buf) \ + { \ + return sprintf(buf, format, ##args); \ + } \ + static DEVICE_ATTR(name, 0444, show_##name, NULL) + +CMM_SHOW(loaned_kb, "%lu\n", PAGES2KB(atomic_long_read(&loaned_pages))); +CMM_SHOW(loaned_target_kb, "%lu\n", PAGES2KB(loaned_pages_target)); + +static ssize_t show_oom_pages(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", PAGES2KB(oom_freed_pages)); +} + +static ssize_t store_oom_pages(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + unsigned long val = simple_strtoul (buf, NULL, 10); + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (val != 0) + return -EBADMSG; + + oom_freed_pages = 0; + return count; +} + +static DEVICE_ATTR(oom_freed_kb, 0644, + show_oom_pages, store_oom_pages); + +static struct device_attribute *cmm_attrs[] = { + &dev_attr_loaned_kb, + &dev_attr_loaned_target_kb, + &dev_attr_oom_freed_kb, +}; + +static DEVICE_ULONG_ATTR(simulate_loan_target_kb, 0644, + simulate_loan_target_kb); + +static struct bus_type cmm_subsys = { + .name = "cmm", + .dev_name = "cmm", +}; + +static void cmm_release_device(struct device *dev) +{ +} + +/** + * cmm_sysfs_register - Register with sysfs + * + * Return value: + * 0 on success / other on failure + **/ +static int cmm_sysfs_register(struct device *dev) +{ + int i, rc; + + if ((rc = subsys_system_register(&cmm_subsys, NULL))) + return rc; + + dev->id = 0; + dev->bus = &cmm_subsys; + dev->release = cmm_release_device; + + if ((rc = device_register(dev))) + goto subsys_unregister; + + for (i = 0; i < ARRAY_SIZE(cmm_attrs); i++) { + if ((rc = device_create_file(dev, cmm_attrs[i]))) + goto fail; + } + + if (!simulate) + return 0; + rc = device_create_file(dev, &dev_attr_simulate_loan_target_kb.attr); + if (rc) + goto fail; + return 0; + +fail: + while (--i >= 0) + device_remove_file(dev, cmm_attrs[i]); + device_unregister(dev); +subsys_unregister: + bus_unregister(&cmm_subsys); + return rc; +} + +/** + * cmm_unregister_sysfs - Unregister from sysfs + * + **/ +static void cmm_unregister_sysfs(struct device *dev) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(cmm_attrs); i++) + device_remove_file(dev, cmm_attrs[i]); + device_unregister(dev); + bus_unregister(&cmm_subsys); +} + +/** + * cmm_reboot_notifier - Make sure pages are not still marked as "loaned" + * + **/ +static int cmm_reboot_notifier(struct notifier_block *nb, + unsigned long action, void *unused) +{ + if (action == SYS_RESTART) { + if (cmm_thread_ptr) + kthread_stop(cmm_thread_ptr); + cmm_thread_ptr = NULL; + cmm_free_pages(atomic_long_read(&loaned_pages)); + } + return NOTIFY_DONE; +} + +static struct notifier_block cmm_reboot_nb = { + .notifier_call = cmm_reboot_notifier, +}; + +/** + * cmm_memory_cb - Handle memory hotplug notifier calls + * @self: notifier block struct + * @action: action to take + * @arg: struct memory_notify data for handler + * + * Return value: + * NOTIFY_OK or notifier error based on subfunction return value + * + **/ +static int cmm_memory_cb(struct notifier_block *self, + unsigned long action, void *arg) +{ + switch (action) { + case MEM_GOING_OFFLINE: + mutex_lock(&hotplug_mutex); + hotplug_occurred = 1; + break; + case MEM_OFFLINE: + case MEM_CANCEL_OFFLINE: + mutex_unlock(&hotplug_mutex); + cmm_dbg("Memory offline operation complete.\n"); + break; + case MEM_GOING_ONLINE: + case MEM_ONLINE: + case MEM_CANCEL_ONLINE: + break; + } + + return NOTIFY_OK; +} + +static struct notifier_block cmm_mem_nb = { + .notifier_call = cmm_memory_cb, + .priority = CMM_MEM_HOTPLUG_PRI +}; + +#ifdef CONFIG_BALLOON_COMPACTION +static int cmm_migratepage(struct balloon_dev_info *b_dev_info, + struct page *newpage, struct page *page, + enum migrate_mode mode) +{ + unsigned long flags; + + /* + * loan/"inflate" the newpage first. + * + * We might race against the cmm_thread who might discover after our + * loan request that another page is to be unloaned. However, once + * the cmm_thread runs again later, this error will automatically + * be corrected. + */ + if (plpar_page_set_loaned(newpage)) { + /* Unlikely, but possible. Tell the caller not to retry now. */ + pr_err_ratelimited("%s: Cannot set page to loaned.", __func__); + return -EBUSY; + } + + /* balloon page list reference */ + get_page(newpage); + + /* + * When we migrate a page to a different zone, we have to fixup the + * count of both involved zones as we adjusted the managed page count + * when inflating. + */ + if (page_zone(page) != page_zone(newpage)) { + adjust_managed_page_count(page, 1); + adjust_managed_page_count(newpage, -1); + } + + spin_lock_irqsave(&b_dev_info->pages_lock, flags); + balloon_page_insert(b_dev_info, newpage); + balloon_page_delete(page); + b_dev_info->isolated_pages--; + spin_unlock_irqrestore(&b_dev_info->pages_lock, flags); + + /* + * activate/"deflate" the old page. We ignore any errors just like the + * other callers. + */ + plpar_page_set_active(page); + + /* balloon page list reference */ + put_page(page); + + return MIGRATEPAGE_SUCCESS; +} + +static void cmm_balloon_compaction_init(void) +{ + balloon_devinfo_init(&b_dev_info); + b_dev_info.migratepage = cmm_migratepage; +} +#else /* CONFIG_BALLOON_COMPACTION */ +static void cmm_balloon_compaction_init(void) +{ +} +#endif /* CONFIG_BALLOON_COMPACTION */ + +/** + * cmm_init - Module initialization + * + * Return value: + * 0 on success / other on failure + **/ +static int cmm_init(void) +{ + int rc; + + if (!firmware_has_feature(FW_FEATURE_CMO) && !simulate) + return -EOPNOTSUPP; + + cmm_balloon_compaction_init(); + + rc = register_oom_notifier(&cmm_oom_nb); + if (rc < 0) + goto out_balloon_compaction; + + if ((rc = register_reboot_notifier(&cmm_reboot_nb))) + goto out_oom_notifier; + + if ((rc = cmm_sysfs_register(&cmm_dev))) + goto out_reboot_notifier; + + rc = register_memory_notifier(&cmm_mem_nb); + if (rc) + goto out_unregister_notifier; + + if (cmm_disabled) + return 0; + + cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread"); + if (IS_ERR(cmm_thread_ptr)) { + rc = PTR_ERR(cmm_thread_ptr); + goto out_unregister_notifier; + } + + return 0; +out_unregister_notifier: + unregister_memory_notifier(&cmm_mem_nb); + cmm_unregister_sysfs(&cmm_dev); +out_reboot_notifier: + unregister_reboot_notifier(&cmm_reboot_nb); +out_oom_notifier: + unregister_oom_notifier(&cmm_oom_nb); +out_balloon_compaction: + return rc; +} + +/** + * cmm_exit - Module exit + * + * Return value: + * nothing + **/ +static void cmm_exit(void) +{ + if (cmm_thread_ptr) + kthread_stop(cmm_thread_ptr); + unregister_oom_notifier(&cmm_oom_nb); + unregister_reboot_notifier(&cmm_reboot_nb); + unregister_memory_notifier(&cmm_mem_nb); + cmm_free_pages(atomic_long_read(&loaned_pages)); + cmm_unregister_sysfs(&cmm_dev); +} + +/** + * cmm_set_disable - Disable/Enable CMM + * + * Return value: + * 0 on success / other on failure + **/ +static int cmm_set_disable(const char *val, const struct kernel_param *kp) +{ + int disable = simple_strtoul(val, NULL, 10); + + if (disable != 0 && disable != 1) + return -EINVAL; + + if (disable && !cmm_disabled) { + if (cmm_thread_ptr) + kthread_stop(cmm_thread_ptr); + cmm_thread_ptr = NULL; + cmm_free_pages(atomic_long_read(&loaned_pages)); + } else if (!disable && cmm_disabled) { + cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread"); + if (IS_ERR(cmm_thread_ptr)) + return PTR_ERR(cmm_thread_ptr); + } + + cmm_disabled = disable; + return 0; +} + +module_param_call(disable, cmm_set_disable, param_get_uint, + &cmm_disabled, 0644); +MODULE_PARM_DESC(disable, "Disable CMM. Set to 1 to disable. " + "[Default=" __stringify(CMM_DISABLE) "]"); + +module_init(cmm_init); +module_exit(cmm_exit); diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c new file mode 100644 index 000000000..47f8eabd1 --- /dev/null +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -0,0 +1,583 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Support for dynamic reconfiguration for PCI, Memory, and CPU + * Hotplug and Dynamic Logical Partitioning on RPA platforms. + * + * Copyright (C) 2009 Nathan Fontenot + * Copyright (C) 2009 IBM Corporation + */ + +#define pr_fmt(fmt) "dlpar: " fmt + +#include <linux/kernel.h> +#include <linux/notifier.h> +#include <linux/spinlock.h> +#include <linux/cpu.h> +#include <linux/slab.h> +#include <linux/of.h> + +#include "of_helpers.h" +#include "pseries.h" + +#include <asm/machdep.h> +#include <linux/uaccess.h> +#include <asm/rtas.h> +#include <asm/rtas-work-area.h> + +static struct workqueue_struct *pseries_hp_wq; + +struct pseries_hp_work { + struct work_struct work; + struct pseries_hp_errorlog *errlog; +}; + +struct cc_workarea { + __be32 drc_index; + __be32 zero; + __be32 name_offset; + __be32 prop_length; + __be32 prop_offset; +}; + +void dlpar_free_cc_property(struct property *prop) +{ + kfree(prop->name); + kfree(prop->value); + kfree(prop); +} + +static struct property *dlpar_parse_cc_property(struct cc_workarea *ccwa) +{ + struct property *prop; + char *name; + char *value; + + prop = kzalloc(sizeof(*prop), GFP_KERNEL); + if (!prop) + return NULL; + + name = (char *)ccwa + be32_to_cpu(ccwa->name_offset); + prop->name = kstrdup(name, GFP_KERNEL); + if (!prop->name) { + dlpar_free_cc_property(prop); + return NULL; + } + + prop->length = be32_to_cpu(ccwa->prop_length); + value = (char *)ccwa + be32_to_cpu(ccwa->prop_offset); + prop->value = kmemdup(value, prop->length, GFP_KERNEL); + if (!prop->value) { + dlpar_free_cc_property(prop); + return NULL; + } + + return prop; +} + +static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa) +{ + struct device_node *dn; + const char *name; + + dn = kzalloc(sizeof(*dn), GFP_KERNEL); + if (!dn) + return NULL; + + name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); + dn->full_name = kstrdup(name, GFP_KERNEL); + if (!dn->full_name) { + kfree(dn); + return NULL; + } + + of_node_set_flag(dn, OF_DYNAMIC); + of_node_init(dn); + + return dn; +} + +static void dlpar_free_one_cc_node(struct device_node *dn) +{ + struct property *prop; + + while (dn->properties) { + prop = dn->properties; + dn->properties = prop->next; + dlpar_free_cc_property(prop); + } + + kfree(dn->full_name); + kfree(dn); +} + +void dlpar_free_cc_nodes(struct device_node *dn) +{ + if (dn->child) + dlpar_free_cc_nodes(dn->child); + + if (dn->sibling) + dlpar_free_cc_nodes(dn->sibling); + + dlpar_free_one_cc_node(dn); +} + +#define COMPLETE 0 +#define NEXT_SIBLING 1 +#define NEXT_CHILD 2 +#define NEXT_PROPERTY 3 +#define PREV_PARENT 4 +#define MORE_MEMORY 5 +#define ERR_CFG_USE -9003 + +struct device_node *dlpar_configure_connector(__be32 drc_index, + struct device_node *parent) +{ + struct device_node *dn; + struct device_node *first_dn = NULL; + struct device_node *last_dn = NULL; + struct property *property; + struct property *last_property = NULL; + struct cc_workarea *ccwa; + struct rtas_work_area *work_area; + char *data_buf; + int cc_token; + int rc = -1; + + cc_token = rtas_function_token(RTAS_FN_IBM_CONFIGURE_CONNECTOR); + if (cc_token == RTAS_UNKNOWN_SERVICE) + return NULL; + + work_area = rtas_work_area_alloc(SZ_4K); + data_buf = rtas_work_area_raw_buf(work_area); + + ccwa = (struct cc_workarea *)&data_buf[0]; + ccwa->drc_index = drc_index; + ccwa->zero = 0; + + do { + do { + rc = rtas_call(cc_token, 2, 1, NULL, + rtas_work_area_phys(work_area), NULL); + } while (rtas_busy_delay(rc)); + + switch (rc) { + case COMPLETE: + break; + + case NEXT_SIBLING: + dn = dlpar_parse_cc_node(ccwa); + if (!dn) + goto cc_error; + + dn->parent = last_dn->parent; + last_dn->sibling = dn; + last_dn = dn; + break; + + case NEXT_CHILD: + dn = dlpar_parse_cc_node(ccwa); + if (!dn) + goto cc_error; + + if (!first_dn) { + dn->parent = parent; + first_dn = dn; + } else { + dn->parent = last_dn; + if (last_dn) + last_dn->child = dn; + } + + last_dn = dn; + break; + + case NEXT_PROPERTY: + property = dlpar_parse_cc_property(ccwa); + if (!property) + goto cc_error; + + if (!last_dn->properties) + last_dn->properties = property; + else + last_property->next = property; + + last_property = property; + break; + + case PREV_PARENT: + last_dn = last_dn->parent; + break; + + case MORE_MEMORY: + case ERR_CFG_USE: + default: + printk(KERN_ERR "Unexpected Error (%d) " + "returned from configure-connector\n", rc); + goto cc_error; + } + } while (rc); + +cc_error: + rtas_work_area_free(work_area); + + if (rc) { + if (first_dn) + dlpar_free_cc_nodes(first_dn); + + return NULL; + } + + return first_dn; +} + +int dlpar_attach_node(struct device_node *dn, struct device_node *parent) +{ + int rc; + + dn->parent = parent; + + rc = of_attach_node(dn); + if (rc) { + printk(KERN_ERR "Failed to add device node %pOF\n", dn); + return rc; + } + + return 0; +} + +int dlpar_detach_node(struct device_node *dn) +{ + struct device_node *child; + int rc; + + child = of_get_next_child(dn, NULL); + while (child) { + dlpar_detach_node(child); + child = of_get_next_child(dn, child); + } + + rc = of_detach_node(dn); + if (rc) + return rc; + + of_node_put(dn); + + return 0; +} + +#define DR_ENTITY_SENSE 9003 +#define DR_ENTITY_PRESENT 1 +#define DR_ENTITY_UNUSABLE 2 +#define ALLOCATION_STATE 9003 +#define ALLOC_UNUSABLE 0 +#define ALLOC_USABLE 1 +#define ISOLATION_STATE 9001 +#define ISOLATE 0 +#define UNISOLATE 1 + +int dlpar_acquire_drc(u32 drc_index) +{ + int dr_status, rc; + + rc = rtas_get_sensor(DR_ENTITY_SENSE, drc_index, &dr_status); + if (rc || dr_status != DR_ENTITY_UNUSABLE) + return -1; + + rc = rtas_set_indicator(ALLOCATION_STATE, drc_index, ALLOC_USABLE); + if (rc) + return rc; + + rc = rtas_set_indicator(ISOLATION_STATE, drc_index, UNISOLATE); + if (rc) { + rtas_set_indicator(ALLOCATION_STATE, drc_index, ALLOC_UNUSABLE); + return rc; + } + + return 0; +} + +int dlpar_release_drc(u32 drc_index) +{ + int dr_status, rc; + + rc = rtas_get_sensor(DR_ENTITY_SENSE, drc_index, &dr_status); + if (rc || dr_status != DR_ENTITY_PRESENT) + return -1; + + rc = rtas_set_indicator(ISOLATION_STATE, drc_index, ISOLATE); + if (rc) + return rc; + + rc = rtas_set_indicator(ALLOCATION_STATE, drc_index, ALLOC_UNUSABLE); + if (rc) { + rtas_set_indicator(ISOLATION_STATE, drc_index, UNISOLATE); + return rc; + } + + return 0; +} + +int dlpar_unisolate_drc(u32 drc_index) +{ + int dr_status, rc; + + rc = rtas_get_sensor(DR_ENTITY_SENSE, drc_index, &dr_status); + if (rc || dr_status != DR_ENTITY_PRESENT) + return -1; + + rtas_set_indicator(ISOLATION_STATE, drc_index, UNISOLATE); + + return 0; +} + +int handle_dlpar_errorlog(struct pseries_hp_errorlog *hp_elog) +{ + int rc; + + /* pseries error logs are in BE format, convert to cpu type */ + switch (hp_elog->id_type) { + case PSERIES_HP_ELOG_ID_DRC_COUNT: + hp_elog->_drc_u.drc_count = + be32_to_cpu(hp_elog->_drc_u.drc_count); + break; + case PSERIES_HP_ELOG_ID_DRC_INDEX: + hp_elog->_drc_u.drc_index = + be32_to_cpu(hp_elog->_drc_u.drc_index); + break; + case PSERIES_HP_ELOG_ID_DRC_IC: + hp_elog->_drc_u.ic.count = + be32_to_cpu(hp_elog->_drc_u.ic.count); + hp_elog->_drc_u.ic.index = + be32_to_cpu(hp_elog->_drc_u.ic.index); + } + + switch (hp_elog->resource) { + case PSERIES_HP_ELOG_RESOURCE_MEM: + rc = dlpar_memory(hp_elog); + break; + case PSERIES_HP_ELOG_RESOURCE_CPU: + rc = dlpar_cpu(hp_elog); + break; + case PSERIES_HP_ELOG_RESOURCE_PMEM: + rc = dlpar_hp_pmem(hp_elog); + break; + + default: + pr_warn_ratelimited("Invalid resource (%d) specified\n", + hp_elog->resource); + rc = -EINVAL; + } + + return rc; +} + +static void pseries_hp_work_fn(struct work_struct *work) +{ + struct pseries_hp_work *hp_work = + container_of(work, struct pseries_hp_work, work); + + handle_dlpar_errorlog(hp_work->errlog); + + kfree(hp_work->errlog); + kfree(work); +} + +void queue_hotplug_event(struct pseries_hp_errorlog *hp_errlog) +{ + struct pseries_hp_work *work; + struct pseries_hp_errorlog *hp_errlog_copy; + + hp_errlog_copy = kmemdup(hp_errlog, sizeof(*hp_errlog), GFP_ATOMIC); + if (!hp_errlog_copy) + return; + + work = kmalloc(sizeof(struct pseries_hp_work), GFP_ATOMIC); + if (work) { + INIT_WORK((struct work_struct *)work, pseries_hp_work_fn); + work->errlog = hp_errlog_copy; + queue_work(pseries_hp_wq, (struct work_struct *)work); + } else { + kfree(hp_errlog_copy); + } +} + +static int dlpar_parse_resource(char **cmd, struct pseries_hp_errorlog *hp_elog) +{ + char *arg; + + arg = strsep(cmd, " "); + if (!arg) + return -EINVAL; + + if (sysfs_streq(arg, "memory")) { + hp_elog->resource = PSERIES_HP_ELOG_RESOURCE_MEM; + } else if (sysfs_streq(arg, "cpu")) { + hp_elog->resource = PSERIES_HP_ELOG_RESOURCE_CPU; + } else { + pr_err("Invalid resource specified.\n"); + return -EINVAL; + } + + return 0; +} + +static int dlpar_parse_action(char **cmd, struct pseries_hp_errorlog *hp_elog) +{ + char *arg; + + arg = strsep(cmd, " "); + if (!arg) + return -EINVAL; + + if (sysfs_streq(arg, "add")) { + hp_elog->action = PSERIES_HP_ELOG_ACTION_ADD; + } else if (sysfs_streq(arg, "remove")) { + hp_elog->action = PSERIES_HP_ELOG_ACTION_REMOVE; + } else { + pr_err("Invalid action specified.\n"); + return -EINVAL; + } + + return 0; +} + +static int dlpar_parse_id_type(char **cmd, struct pseries_hp_errorlog *hp_elog) +{ + char *arg; + u32 count, index; + + arg = strsep(cmd, " "); + if (!arg) + return -EINVAL; + + if (sysfs_streq(arg, "indexed-count")) { + hp_elog->id_type = PSERIES_HP_ELOG_ID_DRC_IC; + arg = strsep(cmd, " "); + if (!arg) { + pr_err("No DRC count specified.\n"); + return -EINVAL; + } + + if (kstrtou32(arg, 0, &count)) { + pr_err("Invalid DRC count specified.\n"); + return -EINVAL; + } + + arg = strsep(cmd, " "); + if (!arg) { + pr_err("No DRC Index specified.\n"); + return -EINVAL; + } + + if (kstrtou32(arg, 0, &index)) { + pr_err("Invalid DRC Index specified.\n"); + return -EINVAL; + } + + hp_elog->_drc_u.ic.count = cpu_to_be32(count); + hp_elog->_drc_u.ic.index = cpu_to_be32(index); + } else if (sysfs_streq(arg, "index")) { + hp_elog->id_type = PSERIES_HP_ELOG_ID_DRC_INDEX; + arg = strsep(cmd, " "); + if (!arg) { + pr_err("No DRC Index specified.\n"); + return -EINVAL; + } + + if (kstrtou32(arg, 0, &index)) { + pr_err("Invalid DRC Index specified.\n"); + return -EINVAL; + } + + hp_elog->_drc_u.drc_index = cpu_to_be32(index); + } else if (sysfs_streq(arg, "count")) { + hp_elog->id_type = PSERIES_HP_ELOG_ID_DRC_COUNT; + arg = strsep(cmd, " "); + if (!arg) { + pr_err("No DRC count specified.\n"); + return -EINVAL; + } + + if (kstrtou32(arg, 0, &count)) { + pr_err("Invalid DRC count specified.\n"); + return -EINVAL; + } + + hp_elog->_drc_u.drc_count = cpu_to_be32(count); + } else { + pr_err("Invalid id_type specified.\n"); + return -EINVAL; + } + + return 0; +} + +static ssize_t dlpar_store(const struct class *class, const struct class_attribute *attr, + const char *buf, size_t count) +{ + struct pseries_hp_errorlog hp_elog; + char *argbuf; + char *args; + int rc; + + args = argbuf = kstrdup(buf, GFP_KERNEL); + if (!argbuf) + return -ENOMEM; + + /* + * Parse out the request from the user, this will be in the form: + * <resource> <action> <id_type> <id> + */ + rc = dlpar_parse_resource(&args, &hp_elog); + if (rc) + goto dlpar_store_out; + + rc = dlpar_parse_action(&args, &hp_elog); + if (rc) + goto dlpar_store_out; + + rc = dlpar_parse_id_type(&args, &hp_elog); + if (rc) + goto dlpar_store_out; + + rc = handle_dlpar_errorlog(&hp_elog); + +dlpar_store_out: + kfree(argbuf); + + if (rc) + pr_err("Could not handle DLPAR request \"%s\"\n", buf); + + return rc ? rc : count; +} + +static ssize_t dlpar_show(const struct class *class, const struct class_attribute *attr, + char *buf) +{ + return sprintf(buf, "%s\n", "memory,cpu"); +} + +static CLASS_ATTR_RW(dlpar); + +int __init dlpar_workqueue_init(void) +{ + if (pseries_hp_wq) + return 0; + + pseries_hp_wq = alloc_ordered_workqueue("pseries hotplug workqueue", 0); + + return pseries_hp_wq ? 0 : -ENOMEM; +} + +static int __init dlpar_sysfs_init(void) +{ + int rc; + + rc = dlpar_workqueue_init(); + if (rc) + return rc; + + return sysfs_create_file(kernel_kobj, &class_attr_dlpar.attr); +} +machine_device_initcall(pseries, dlpar_sysfs_init); + diff --git a/arch/powerpc/platforms/pseries/dtl.c b/arch/powerpc/platforms/pseries/dtl.c new file mode 100644 index 000000000..3f1cdcceb --- /dev/null +++ b/arch/powerpc/platforms/pseries/dtl.c @@ -0,0 +1,445 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Virtual Processor Dispatch Trace Log + * + * (C) Copyright IBM Corporation 2009 + * + * Author: Jeremy Kerr <jk@ozlabs.org> + */ + +#include <linux/slab.h> +#include <linux/spinlock.h> +#include <asm/smp.h> +#include <linux/uaccess.h> +#include <linux/debugfs.h> +#include <asm/firmware.h> +#include <asm/dtl.h> +#include <asm/lppaca.h> +#include <asm/plpar_wrappers.h> +#include <asm/machdep.h> + +#ifdef CONFIG_DTL +struct dtl { + struct dtl_entry *buf; + int cpu; + int buf_entries; + u64 last_idx; + spinlock_t lock; +}; +static DEFINE_PER_CPU(struct dtl, cpu_dtl); + +static u8 dtl_event_mask = DTL_LOG_ALL; + + +/* + * Size of per-cpu log buffers. Firmware requires that the buffer does + * not cross a 4k boundary. + */ +static int dtl_buf_entries = N_DISPATCH_LOG; + +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE + +/* + * When CONFIG_VIRT_CPU_ACCOUNTING_NATIVE = y, the cpu accounting code controls + * reading from the dispatch trace log. If other code wants to consume + * DTL entries, it can set this pointer to a function that will get + * called once for each DTL entry that gets processed. + */ +static void (*dtl_consumer)(struct dtl_entry *entry, u64 index); + +struct dtl_ring { + u64 write_index; + struct dtl_entry *write_ptr; + struct dtl_entry *buf; + struct dtl_entry *buf_end; +}; + +static DEFINE_PER_CPU(struct dtl_ring, dtl_rings); + +static atomic_t dtl_count; + +/* + * The cpu accounting code controls the DTL ring buffer, and we get + * given entries as they are processed. + */ +static void consume_dtle(struct dtl_entry *dtle, u64 index) +{ + struct dtl_ring *dtlr = this_cpu_ptr(&dtl_rings); + struct dtl_entry *wp = dtlr->write_ptr; + struct lppaca *vpa = local_paca->lppaca_ptr; + + if (!wp) + return; + + *wp = *dtle; + barrier(); + + /* check for hypervisor ring buffer overflow, ignore this entry if so */ + if (index + N_DISPATCH_LOG < be64_to_cpu(vpa->dtl_idx)) + return; + + ++wp; + if (wp == dtlr->buf_end) + wp = dtlr->buf; + dtlr->write_ptr = wp; + + /* incrementing write_index makes the new entry visible */ + smp_wmb(); + ++dtlr->write_index; +} + +static int dtl_start(struct dtl *dtl) +{ + struct dtl_ring *dtlr = &per_cpu(dtl_rings, dtl->cpu); + + dtlr->buf = dtl->buf; + dtlr->buf_end = dtl->buf + dtl->buf_entries; + dtlr->write_index = 0; + + /* setting write_ptr enables logging into our buffer */ + smp_wmb(); + dtlr->write_ptr = dtl->buf; + + /* enable event logging */ + lppaca_of(dtl->cpu).dtl_enable_mask |= dtl_event_mask; + + dtl_consumer = consume_dtle; + atomic_inc(&dtl_count); + return 0; +} + +static void dtl_stop(struct dtl *dtl) +{ + struct dtl_ring *dtlr = &per_cpu(dtl_rings, dtl->cpu); + + dtlr->write_ptr = NULL; + smp_wmb(); + + dtlr->buf = NULL; + + /* restore dtl_enable_mask */ + lppaca_of(dtl->cpu).dtl_enable_mask = DTL_LOG_PREEMPT; + + if (atomic_dec_and_test(&dtl_count)) + dtl_consumer = NULL; +} + +static u64 dtl_current_index(struct dtl *dtl) +{ + return per_cpu(dtl_rings, dtl->cpu).write_index; +} + +#else /* CONFIG_VIRT_CPU_ACCOUNTING_NATIVE */ + +static int dtl_start(struct dtl *dtl) +{ + unsigned long addr; + int ret, hwcpu; + + /* Register our dtl buffer with the hypervisor. The HV expects the + * buffer size to be passed in the second word of the buffer */ + ((u32 *)dtl->buf)[1] = cpu_to_be32(DISPATCH_LOG_BYTES); + + hwcpu = get_hard_smp_processor_id(dtl->cpu); + addr = __pa(dtl->buf); + ret = register_dtl(hwcpu, addr); + if (ret) { + printk(KERN_WARNING "%s: DTL registration for cpu %d (hw %d) " + "failed with %d\n", __func__, dtl->cpu, hwcpu, ret); + return -EIO; + } + + /* set our initial buffer indices */ + lppaca_of(dtl->cpu).dtl_idx = 0; + + /* ensure that our updates to the lppaca fields have occurred before + * we actually enable the logging */ + smp_wmb(); + + /* enable event logging */ + lppaca_of(dtl->cpu).dtl_enable_mask = dtl_event_mask; + + return 0; +} + +static void dtl_stop(struct dtl *dtl) +{ + int hwcpu = get_hard_smp_processor_id(dtl->cpu); + + lppaca_of(dtl->cpu).dtl_enable_mask = 0x0; + + unregister_dtl(hwcpu); +} + +static u64 dtl_current_index(struct dtl *dtl) +{ + return be64_to_cpu(lppaca_of(dtl->cpu).dtl_idx); +} +#endif /* CONFIG_VIRT_CPU_ACCOUNTING_NATIVE */ + +static int dtl_enable(struct dtl *dtl) +{ + long int n_entries; + long int rc; + struct dtl_entry *buf = NULL; + + if (!dtl_cache) + return -ENOMEM; + + /* only allow one reader */ + if (dtl->buf) + return -EBUSY; + + /* ensure there are no other conflicting dtl users */ + if (!read_trylock(&dtl_access_lock)) + return -EBUSY; + + n_entries = dtl_buf_entries; + buf = kmem_cache_alloc_node(dtl_cache, GFP_KERNEL, cpu_to_node(dtl->cpu)); + if (!buf) { + printk(KERN_WARNING "%s: buffer alloc failed for cpu %d\n", + __func__, dtl->cpu); + read_unlock(&dtl_access_lock); + return -ENOMEM; + } + + spin_lock(&dtl->lock); + rc = -EBUSY; + if (!dtl->buf) { + /* store the original allocation size for use during read */ + dtl->buf_entries = n_entries; + dtl->buf = buf; + dtl->last_idx = 0; + rc = dtl_start(dtl); + if (rc) + dtl->buf = NULL; + } + spin_unlock(&dtl->lock); + + if (rc) { + read_unlock(&dtl_access_lock); + kmem_cache_free(dtl_cache, buf); + } + + return rc; +} + +static void dtl_disable(struct dtl *dtl) +{ + spin_lock(&dtl->lock); + dtl_stop(dtl); + kmem_cache_free(dtl_cache, dtl->buf); + dtl->buf = NULL; + dtl->buf_entries = 0; + spin_unlock(&dtl->lock); + read_unlock(&dtl_access_lock); +} + +/* file interface */ + +static int dtl_file_open(struct inode *inode, struct file *filp) +{ + struct dtl *dtl = inode->i_private; + int rc; + + rc = dtl_enable(dtl); + if (rc) + return rc; + + filp->private_data = dtl; + return 0; +} + +static int dtl_file_release(struct inode *inode, struct file *filp) +{ + struct dtl *dtl = inode->i_private; + dtl_disable(dtl); + return 0; +} + +static ssize_t dtl_file_read(struct file *filp, char __user *buf, size_t len, + loff_t *pos) +{ + long int rc, n_read, n_req, read_size; + struct dtl *dtl; + u64 cur_idx, last_idx, i; + + if ((len % sizeof(struct dtl_entry)) != 0) + return -EINVAL; + + dtl = filp->private_data; + + /* requested number of entries to read */ + n_req = len / sizeof(struct dtl_entry); + + /* actual number of entries read */ + n_read = 0; + + spin_lock(&dtl->lock); + + cur_idx = dtl_current_index(dtl); + last_idx = dtl->last_idx; + + if (last_idx + dtl->buf_entries <= cur_idx) + last_idx = cur_idx - dtl->buf_entries + 1; + + if (last_idx + n_req > cur_idx) + n_req = cur_idx - last_idx; + + if (n_req > 0) + dtl->last_idx = last_idx + n_req; + + spin_unlock(&dtl->lock); + + if (n_req <= 0) + return 0; + + i = last_idx % dtl->buf_entries; + + /* read the tail of the buffer if we've wrapped */ + if (i + n_req > dtl->buf_entries) { + read_size = dtl->buf_entries - i; + + rc = copy_to_user(buf, &dtl->buf[i], + read_size * sizeof(struct dtl_entry)); + if (rc) + return -EFAULT; + + i = 0; + n_req -= read_size; + n_read += read_size; + buf += read_size * sizeof(struct dtl_entry); + } + + /* .. and now the head */ + rc = copy_to_user(buf, &dtl->buf[i], n_req * sizeof(struct dtl_entry)); + if (rc) + return -EFAULT; + + n_read += n_req; + + return n_read * sizeof(struct dtl_entry); +} + +static const struct file_operations dtl_fops = { + .open = dtl_file_open, + .release = dtl_file_release, + .read = dtl_file_read, + .llseek = no_llseek, +}; + +static struct dentry *dtl_dir; + +static void dtl_setup_file(struct dtl *dtl) +{ + char name[10]; + + sprintf(name, "cpu-%d", dtl->cpu); + + debugfs_create_file(name, 0400, dtl_dir, dtl, &dtl_fops); +} + +static int dtl_init(void) +{ + int i; + + if (!firmware_has_feature(FW_FEATURE_SPLPAR)) + return -ENODEV; + + /* set up common debugfs structure */ + + dtl_dir = debugfs_create_dir("dtl", arch_debugfs_dir); + + debugfs_create_x8("dtl_event_mask", 0600, dtl_dir, &dtl_event_mask); + debugfs_create_u32("dtl_buf_entries", 0400, dtl_dir, &dtl_buf_entries); + + /* set up the per-cpu log structures */ + for_each_possible_cpu(i) { + struct dtl *dtl = &per_cpu(cpu_dtl, i); + spin_lock_init(&dtl->lock); + dtl->cpu = i; + + dtl_setup_file(dtl); + } + + return 0; +} +machine_arch_initcall(pseries, dtl_init); +#endif /* CONFIG_DTL */ + +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE +/* + * Scan the dispatch trace log and count up the stolen time. + * Should be called with interrupts disabled. + */ +static notrace u64 scan_dispatch_log(u64 stop_tb) +{ + u64 i = local_paca->dtl_ridx; + struct dtl_entry *dtl = local_paca->dtl_curr; + struct dtl_entry *dtl_end = local_paca->dispatch_log_end; + struct lppaca *vpa = local_paca->lppaca_ptr; + u64 tb_delta; + u64 stolen = 0; + u64 dtb; + + if (!dtl) + return 0; + + if (i == be64_to_cpu(vpa->dtl_idx)) + return 0; + while (i < be64_to_cpu(vpa->dtl_idx)) { + dtb = be64_to_cpu(dtl->timebase); + tb_delta = be32_to_cpu(dtl->enqueue_to_dispatch_time) + + be32_to_cpu(dtl->ready_to_enqueue_time); + barrier(); + if (i + N_DISPATCH_LOG < be64_to_cpu(vpa->dtl_idx)) { + /* buffer has overflowed */ + i = be64_to_cpu(vpa->dtl_idx) - N_DISPATCH_LOG; + dtl = local_paca->dispatch_log + (i % N_DISPATCH_LOG); + continue; + } + if (dtb > stop_tb) + break; +#ifdef CONFIG_DTL + if (dtl_consumer) + dtl_consumer(dtl, i); +#endif + stolen += tb_delta; + ++i; + ++dtl; + if (dtl == dtl_end) + dtl = local_paca->dispatch_log; + } + local_paca->dtl_ridx = i; + local_paca->dtl_curr = dtl; + return stolen; +} + +/* + * Accumulate stolen time by scanning the dispatch trace log. + * Called on entry from user mode. + */ +void notrace pseries_accumulate_stolen_time(void) +{ + u64 sst, ust; + struct cpu_accounting_data *acct = &local_paca->accounting; + + sst = scan_dispatch_log(acct->starttime_user); + ust = scan_dispatch_log(acct->starttime); + acct->stime -= sst; + acct->utime -= ust; + acct->steal_time += ust + sst; +} + +u64 pseries_calculate_stolen_time(u64 stop_tb) +{ + if (!firmware_has_feature(FW_FEATURE_SPLPAR)) + return 0; + + if (get_paca()->dtl_ridx != be64_to_cpu(get_lppaca()->dtl_idx)) + return scan_dispatch_log(stop_tb); + + return 0; +} + +#endif diff --git a/arch/powerpc/platforms/pseries/eeh_pseries.c b/arch/powerpc/platforms/pseries/eeh_pseries.c new file mode 100644 index 000000000..def184da5 --- /dev/null +++ b/arch/powerpc/platforms/pseries/eeh_pseries.c @@ -0,0 +1,887 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * The file intends to implement the platform dependent EEH operations on pseries. + * Actually, the pseries platform is built based on RTAS heavily. That means the + * pseries platform dependent EEH operations will be built on RTAS calls. The functions + * are derived from arch/powerpc/platforms/pseries/eeh.c and necessary cleanup has + * been done. + * + * Copyright Benjamin Herrenschmidt & Gavin Shan, IBM Corporation 2011. + * Copyright IBM Corporation 2001, 2005, 2006 + * Copyright Dave Engebretsen & Todd Inglett 2001 + * Copyright Linas Vepstas 2005, 2006 + */ + +#include <linux/atomic.h> +#include <linux/delay.h> +#include <linux/export.h> +#include <linux/init.h> +#include <linux/list.h> +#include <linux/of.h> +#include <linux/pci.h> +#include <linux/proc_fs.h> +#include <linux/rbtree.h> +#include <linux/sched.h> +#include <linux/seq_file.h> +#include <linux/spinlock.h> +#include <linux/crash_dump.h> + +#include <asm/eeh.h> +#include <asm/eeh_event.h> +#include <asm/io.h> +#include <asm/machdep.h> +#include <asm/ppc-pci.h> +#include <asm/rtas.h> + +/* RTAS tokens */ +static int ibm_set_eeh_option; +static int ibm_set_slot_reset; +static int ibm_read_slot_reset_state; +static int ibm_read_slot_reset_state2; +static int ibm_slot_error_detail; +static int ibm_get_config_addr_info; +static int ibm_get_config_addr_info2; +static int ibm_configure_pe; + +static void pseries_eeh_init_edev(struct pci_dn *pdn); + +static void pseries_pcibios_bus_add_device(struct pci_dev *pdev) +{ + struct pci_dn *pdn = pci_get_pdn(pdev); + + if (eeh_has_flag(EEH_FORCE_DISABLED)) + return; + + dev_dbg(&pdev->dev, "EEH: Setting up device\n"); +#ifdef CONFIG_PCI_IOV + if (pdev->is_virtfn) { + pdn->device_id = pdev->device; + pdn->vendor_id = pdev->vendor; + pdn->class_code = pdev->class; + /* + * Last allow unfreeze return code used for retrieval + * by user space in eeh-sysfs to show the last command + * completion from platform. + */ + pdn->last_allow_rc = 0; + } +#endif + pseries_eeh_init_edev(pdn); +#ifdef CONFIG_PCI_IOV + if (pdev->is_virtfn) { + /* + * FIXME: This really should be handled by choosing the right + * parent PE in pseries_eeh_init_edev(). + */ + struct eeh_pe *physfn_pe = pci_dev_to_eeh_dev(pdev->physfn)->pe; + struct eeh_dev *edev = pdn_to_eeh_dev(pdn); + + edev->pe_config_addr = (pdn->busno << 16) | (pdn->devfn << 8); + eeh_pe_tree_remove(edev); /* Remove as it is adding to bus pe */ + eeh_pe_tree_insert(edev, physfn_pe); /* Add as VF PE type */ + } +#endif + eeh_probe_device(pdev); +} + + +/** + * pseries_eeh_get_pe_config_addr - Find the pe_config_addr for a device + * @pdn: pci_dn of the input device + * + * The EEH RTAS calls use a tuple consisting of: (buid_hi, buid_lo, + * pe_config_addr) as a handle to a given PE. This function finds the + * pe_config_addr based on the device's config addr. + * + * Keep in mind that the pe_config_addr *might* be numerically identical to the + * device's config addr, but the two are conceptually distinct. + * + * Returns the pe_config_addr, or a negative error code. + */ +static int pseries_eeh_get_pe_config_addr(struct pci_dn *pdn) +{ + int config_addr = rtas_config_addr(pdn->busno, pdn->devfn, 0); + struct pci_controller *phb = pdn->phb; + int ret, rets[3]; + + if (ibm_get_config_addr_info2 != RTAS_UNKNOWN_SERVICE) { + /* + * First of all, use function 1 to determine if this device is + * part of a PE or not. ret[0] being zero indicates it's not. + */ + ret = rtas_call(ibm_get_config_addr_info2, 4, 2, rets, + config_addr, BUID_HI(phb->buid), + BUID_LO(phb->buid), 1); + if (ret || (rets[0] == 0)) + return -ENOENT; + + /* Retrieve the associated PE config address with function 0 */ + ret = rtas_call(ibm_get_config_addr_info2, 4, 2, rets, + config_addr, BUID_HI(phb->buid), + BUID_LO(phb->buid), 0); + if (ret) { + pr_warn("%s: Failed to get address for PHB#%x-PE#%x\n", + __func__, phb->global_number, config_addr); + return -ENXIO; + } + + return rets[0]; + } + + if (ibm_get_config_addr_info != RTAS_UNKNOWN_SERVICE) { + ret = rtas_call(ibm_get_config_addr_info, 4, 2, rets, + config_addr, BUID_HI(phb->buid), + BUID_LO(phb->buid), 0); + if (ret) { + pr_warn("%s: Failed to get address for PHB#%x-PE#%x\n", + __func__, phb->global_number, config_addr); + return -ENXIO; + } + + return rets[0]; + } + + /* + * PAPR does describe a process for finding the pe_config_addr that was + * used before the ibm,get-config-addr-info calls were added. However, + * I haven't found *any* systems that don't have that RTAS call + * implemented. If you happen to find one that needs the old DT based + * process, patches are welcome! + */ + return -ENOENT; +} + +/** + * pseries_eeh_phb_reset - Reset the specified PHB + * @phb: PCI controller + * @config_addr: the associated config address + * @option: reset option + * + * Reset the specified PHB/PE + */ +static int pseries_eeh_phb_reset(struct pci_controller *phb, int config_addr, int option) +{ + int ret; + + /* Reset PE through RTAS call */ + ret = rtas_call(ibm_set_slot_reset, 4, 1, NULL, + config_addr, BUID_HI(phb->buid), + BUID_LO(phb->buid), option); + + /* If fundamental-reset not supported, try hot-reset */ + if (option == EEH_RESET_FUNDAMENTAL && ret == -8) { + option = EEH_RESET_HOT; + ret = rtas_call(ibm_set_slot_reset, 4, 1, NULL, + config_addr, BUID_HI(phb->buid), + BUID_LO(phb->buid), option); + } + + /* We need reset hold or settlement delay */ + if (option == EEH_RESET_FUNDAMENTAL || option == EEH_RESET_HOT) + msleep(EEH_PE_RST_HOLD_TIME); + else + msleep(EEH_PE_RST_SETTLE_TIME); + + return ret; +} + +/** + * pseries_eeh_phb_configure_bridge - Configure PCI bridges in the indicated PE + * @phb: PCI controller + * @config_addr: the associated config address + * + * The function will be called to reconfigure the bridges included + * in the specified PE so that the mulfunctional PE would be recovered + * again. + */ +static int pseries_eeh_phb_configure_bridge(struct pci_controller *phb, int config_addr) +{ + int ret; + /* Waiting 0.2s maximum before skipping configuration */ + int max_wait = 200; + + while (max_wait > 0) { + ret = rtas_call(ibm_configure_pe, 3, 1, NULL, + config_addr, BUID_HI(phb->buid), + BUID_LO(phb->buid)); + + if (!ret) + return ret; + if (ret < 0) + break; + + /* + * If RTAS returns a delay value that's above 100ms, cut it + * down to 100ms in case firmware made a mistake. For more + * on how these delay values work see rtas_busy_delay_time + */ + if (ret > RTAS_EXTENDED_DELAY_MIN+2 && + ret <= RTAS_EXTENDED_DELAY_MAX) + ret = RTAS_EXTENDED_DELAY_MIN+2; + + max_wait -= rtas_busy_delay_time(ret); + + if (max_wait < 0) + break; + + rtas_busy_delay(ret); + } + + pr_warn("%s: Unable to configure bridge PHB#%x-PE#%x (%d)\n", + __func__, phb->global_number, config_addr, ret); + /* PAPR defines -3 as "Parameter Error" for this function: */ + if (ret == -3) + return -EINVAL; + else + return -EIO; +} + +/* + * Buffer for reporting slot-error-detail rtas calls. Its here + * in BSS, and not dynamically alloced, so that it ends up in + * RMO where RTAS can access it. + */ +static unsigned char slot_errbuf[RTAS_ERROR_LOG_MAX]; +static DEFINE_SPINLOCK(slot_errbuf_lock); +static int eeh_error_buf_size; + +static int pseries_eeh_cap_start(struct pci_dn *pdn) +{ + u32 status; + + if (!pdn) + return 0; + + rtas_read_config(pdn, PCI_STATUS, 2, &status); + if (!(status & PCI_STATUS_CAP_LIST)) + return 0; + + return PCI_CAPABILITY_LIST; +} + + +static int pseries_eeh_find_cap(struct pci_dn *pdn, int cap) +{ + int pos = pseries_eeh_cap_start(pdn); + int cnt = 48; /* Maximal number of capabilities */ + u32 id; + + if (!pos) + return 0; + + while (cnt--) { + rtas_read_config(pdn, pos, 1, &pos); + if (pos < 0x40) + break; + pos &= ~3; + rtas_read_config(pdn, pos + PCI_CAP_LIST_ID, 1, &id); + if (id == 0xff) + break; + if (id == cap) + return pos; + pos += PCI_CAP_LIST_NEXT; + } + + return 0; +} + +static int pseries_eeh_find_ecap(struct pci_dn *pdn, int cap) +{ + struct eeh_dev *edev = pdn_to_eeh_dev(pdn); + u32 header; + int pos = 256; + int ttl = (4096 - 256) / 8; + + if (!edev || !edev->pcie_cap) + return 0; + if (rtas_read_config(pdn, pos, 4, &header) != PCIBIOS_SUCCESSFUL) + return 0; + else if (!header) + return 0; + + while (ttl-- > 0) { + if (PCI_EXT_CAP_ID(header) == cap && pos) + return pos; + + pos = PCI_EXT_CAP_NEXT(header); + if (pos < 256) + break; + + if (rtas_read_config(pdn, pos, 4, &header) != PCIBIOS_SUCCESSFUL) + break; + } + + return 0; +} + +/** + * pseries_eeh_pe_get_parent - Retrieve the parent PE + * @edev: EEH device + * + * The whole PEs existing in the system are organized as hierarchy + * tree. The function is used to retrieve the parent PE according + * to the parent EEH device. + */ +static struct eeh_pe *pseries_eeh_pe_get_parent(struct eeh_dev *edev) +{ + struct eeh_dev *parent; + struct pci_dn *pdn = eeh_dev_to_pdn(edev); + + /* + * It might have the case for the indirect parent + * EEH device already having associated PE, but + * the direct parent EEH device doesn't have yet. + */ + if (edev->physfn) + pdn = pci_get_pdn(edev->physfn); + else + pdn = pdn ? pdn->parent : NULL; + while (pdn) { + /* We're poking out of PCI territory */ + parent = pdn_to_eeh_dev(pdn); + if (!parent) + return NULL; + + if (parent->pe) + return parent->pe; + + pdn = pdn->parent; + } + + return NULL; +} + +/** + * pseries_eeh_init_edev - initialise the eeh_dev and eeh_pe for a pci_dn + * + * @pdn: PCI device node + * + * When we discover a new PCI device via the device-tree we create a + * corresponding pci_dn and we allocate, but don't initialise, an eeh_dev. + * This function takes care of the initialisation and inserts the eeh_dev + * into the correct eeh_pe. If no eeh_pe exists we'll allocate one. + */ +static void pseries_eeh_init_edev(struct pci_dn *pdn) +{ + struct eeh_pe pe, *parent; + struct eeh_dev *edev; + u32 pcie_flags; + int ret; + + if (WARN_ON_ONCE(!eeh_has_flag(EEH_PROBE_MODE_DEVTREE))) + return; + + /* + * Find the eeh_dev for this pdn. The storage for the eeh_dev was + * allocated at the same time as the pci_dn. + * + * XXX: We should probably re-visit that. + */ + edev = pdn_to_eeh_dev(pdn); + if (!edev) + return; + + /* + * If ->pe is set then we've already probed this device. We hit + * this path when a pci_dev is removed and rescanned while recovering + * a PE (i.e. for devices where the driver doesn't support error + * recovery). + */ + if (edev->pe) + return; + + /* Check class/vendor/device IDs */ + if (!pdn->vendor_id || !pdn->device_id || !pdn->class_code) + return; + + /* Skip for PCI-ISA bridge */ + if ((pdn->class_code >> 8) == PCI_CLASS_BRIDGE_ISA) + return; + + eeh_edev_dbg(edev, "Probing device\n"); + + /* + * Update class code and mode of eeh device. We need + * correctly reflects that current device is root port + * or PCIe switch downstream port. + */ + edev->pcix_cap = pseries_eeh_find_cap(pdn, PCI_CAP_ID_PCIX); + edev->pcie_cap = pseries_eeh_find_cap(pdn, PCI_CAP_ID_EXP); + edev->aer_cap = pseries_eeh_find_ecap(pdn, PCI_EXT_CAP_ID_ERR); + edev->mode &= 0xFFFFFF00; + if ((pdn->class_code >> 8) == PCI_CLASS_BRIDGE_PCI) { + edev->mode |= EEH_DEV_BRIDGE; + if (edev->pcie_cap) { + rtas_read_config(pdn, edev->pcie_cap + PCI_EXP_FLAGS, + 2, &pcie_flags); + pcie_flags = (pcie_flags & PCI_EXP_FLAGS_TYPE) >> 4; + if (pcie_flags == PCI_EXP_TYPE_ROOT_PORT) + edev->mode |= EEH_DEV_ROOT_PORT; + else if (pcie_flags == PCI_EXP_TYPE_DOWNSTREAM) + edev->mode |= EEH_DEV_DS_PORT; + } + } + + /* first up, find the pe_config_addr for the PE containing the device */ + ret = pseries_eeh_get_pe_config_addr(pdn); + if (ret < 0) { + eeh_edev_dbg(edev, "Unable to find pe_config_addr\n"); + goto err; + } + + /* Try enable EEH on the fake PE */ + memset(&pe, 0, sizeof(struct eeh_pe)); + pe.phb = pdn->phb; + pe.addr = ret; + + eeh_edev_dbg(edev, "Enabling EEH on device\n"); + ret = eeh_ops->set_option(&pe, EEH_OPT_ENABLE); + if (ret) { + eeh_edev_dbg(edev, "EEH failed to enable on device (code %d)\n", ret); + goto err; + } + + edev->pe_config_addr = pe.addr; + + eeh_add_flag(EEH_ENABLED); + + parent = pseries_eeh_pe_get_parent(edev); + eeh_pe_tree_insert(edev, parent); + eeh_save_bars(edev); + eeh_edev_dbg(edev, "EEH enabled for device"); + + return; + +err: + eeh_edev_dbg(edev, "EEH is unsupported on device (code = %d)\n", ret); +} + +static struct eeh_dev *pseries_eeh_probe(struct pci_dev *pdev) +{ + struct eeh_dev *edev; + struct pci_dn *pdn; + + pdn = pci_get_pdn_by_devfn(pdev->bus, pdev->devfn); + if (!pdn) + return NULL; + + /* + * If the system supports EEH on this device then the eeh_dev was + * configured and inserted into a PE in pseries_eeh_init_edev() + */ + edev = pdn_to_eeh_dev(pdn); + if (!edev || !edev->pe) + return NULL; + + return edev; +} + +/** + * pseries_eeh_init_edev_recursive - Enable EEH for the indicated device + * @pdn: PCI device node + * + * This routine must be used to perform EEH initialization for the + * indicated PCI device that was added after system boot (e.g. + * hotplug, dlpar). + */ +void pseries_eeh_init_edev_recursive(struct pci_dn *pdn) +{ + struct pci_dn *n; + + if (!pdn) + return; + + list_for_each_entry(n, &pdn->child_list, list) + pseries_eeh_init_edev_recursive(n); + + pseries_eeh_init_edev(pdn); +} +EXPORT_SYMBOL_GPL(pseries_eeh_init_edev_recursive); + +/** + * pseries_eeh_set_option - Initialize EEH or MMIO/DMA reenable + * @pe: EEH PE + * @option: operation to be issued + * + * The function is used to control the EEH functionality globally. + * Currently, following options are support according to PAPR: + * Enable EEH, Disable EEH, Enable MMIO and Enable DMA + */ +static int pseries_eeh_set_option(struct eeh_pe *pe, int option) +{ + int ret = 0; + + /* + * When we're enabling or disabling EEH functionality on + * the particular PE, the PE config address is possibly + * unavailable. Therefore, we have to figure it out from + * the FDT node. + */ + switch (option) { + case EEH_OPT_DISABLE: + case EEH_OPT_ENABLE: + case EEH_OPT_THAW_MMIO: + case EEH_OPT_THAW_DMA: + break; + case EEH_OPT_FREEZE_PE: + /* Not support */ + return 0; + default: + pr_err("%s: Invalid option %d\n", __func__, option); + return -EINVAL; + } + + ret = rtas_call(ibm_set_eeh_option, 4, 1, NULL, + pe->addr, BUID_HI(pe->phb->buid), + BUID_LO(pe->phb->buid), option); + + return ret; +} + +/** + * pseries_eeh_get_state - Retrieve PE state + * @pe: EEH PE + * @delay: suggested time to wait if state is unavailable + * + * Retrieve the state of the specified PE. On RTAS compliant + * pseries platform, there already has one dedicated RTAS function + * for the purpose. It's notable that the associated PE config address + * might be ready when calling the function. Therefore, endeavour to + * use the PE config address if possible. Further more, there're 2 + * RTAS calls for the purpose, we need to try the new one and back + * to the old one if the new one couldn't work properly. + */ +static int pseries_eeh_get_state(struct eeh_pe *pe, int *delay) +{ + int ret; + int rets[4]; + int result; + + if (ibm_read_slot_reset_state2 != RTAS_UNKNOWN_SERVICE) { + ret = rtas_call(ibm_read_slot_reset_state2, 3, 4, rets, + pe->addr, BUID_HI(pe->phb->buid), + BUID_LO(pe->phb->buid)); + } else if (ibm_read_slot_reset_state != RTAS_UNKNOWN_SERVICE) { + /* Fake PE unavailable info */ + rets[2] = 0; + ret = rtas_call(ibm_read_slot_reset_state, 3, 3, rets, + pe->addr, BUID_HI(pe->phb->buid), + BUID_LO(pe->phb->buid)); + } else { + return EEH_STATE_NOT_SUPPORT; + } + + if (ret) + return ret; + + /* Parse the result out */ + if (!rets[1]) + return EEH_STATE_NOT_SUPPORT; + + switch(rets[0]) { + case 0: + result = EEH_STATE_MMIO_ACTIVE | + EEH_STATE_DMA_ACTIVE; + break; + case 1: + result = EEH_STATE_RESET_ACTIVE | + EEH_STATE_MMIO_ACTIVE | + EEH_STATE_DMA_ACTIVE; + break; + case 2: + result = 0; + break; + case 4: + result = EEH_STATE_MMIO_ENABLED; + break; + case 5: + if (rets[2]) { + if (delay) + *delay = rets[2]; + result = EEH_STATE_UNAVAILABLE; + } else { + result = EEH_STATE_NOT_SUPPORT; + } + break; + default: + result = EEH_STATE_NOT_SUPPORT; + } + + return result; +} + +/** + * pseries_eeh_reset - Reset the specified PE + * @pe: EEH PE + * @option: reset option + * + * Reset the specified PE + */ +static int pseries_eeh_reset(struct eeh_pe *pe, int option) +{ + return pseries_eeh_phb_reset(pe->phb, pe->addr, option); +} + +/** + * pseries_eeh_get_log - Retrieve error log + * @pe: EEH PE + * @severity: temporary or permanent error log + * @drv_log: driver log to be combined with retrieved error log + * @len: length of driver log + * + * Retrieve the temporary or permanent error from the PE. + * Actually, the error will be retrieved through the dedicated + * RTAS call. + */ +static int pseries_eeh_get_log(struct eeh_pe *pe, int severity, char *drv_log, unsigned long len) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&slot_errbuf_lock, flags); + memset(slot_errbuf, 0, eeh_error_buf_size); + + ret = rtas_call(ibm_slot_error_detail, 8, 1, NULL, pe->addr, + BUID_HI(pe->phb->buid), BUID_LO(pe->phb->buid), + virt_to_phys(drv_log), len, + virt_to_phys(slot_errbuf), eeh_error_buf_size, + severity); + if (!ret) + log_error(slot_errbuf, ERR_TYPE_RTAS_LOG, 0); + spin_unlock_irqrestore(&slot_errbuf_lock, flags); + + return ret; +} + +/** + * pseries_eeh_configure_bridge - Configure PCI bridges in the indicated PE + * @pe: EEH PE + * + */ +static int pseries_eeh_configure_bridge(struct eeh_pe *pe) +{ + return pseries_eeh_phb_configure_bridge(pe->phb, pe->addr); +} + +/** + * pseries_eeh_read_config - Read PCI config space + * @edev: EEH device handle + * @where: PCI config space offset + * @size: size to read + * @val: return value + * + * Read config space from the speicifed device + */ +static int pseries_eeh_read_config(struct eeh_dev *edev, int where, int size, u32 *val) +{ + struct pci_dn *pdn = eeh_dev_to_pdn(edev); + + return rtas_read_config(pdn, where, size, val); +} + +/** + * pseries_eeh_write_config - Write PCI config space + * @edev: EEH device handle + * @where: PCI config space offset + * @size: size to write + * @val: value to be written + * + * Write config space to the specified device + */ +static int pseries_eeh_write_config(struct eeh_dev *edev, int where, int size, u32 val) +{ + struct pci_dn *pdn = eeh_dev_to_pdn(edev); + + return rtas_write_config(pdn, where, size, val); +} + +#ifdef CONFIG_PCI_IOV +static int pseries_send_allow_unfreeze(struct pci_dn *pdn, u16 *vf_pe_array, int cur_vfs) +{ + int rc; + int ibm_allow_unfreeze = rtas_function_token(RTAS_FN_IBM_OPEN_SRIOV_ALLOW_UNFREEZE); + unsigned long buid, addr; + + addr = rtas_config_addr(pdn->busno, pdn->devfn, 0); + buid = pdn->phb->buid; + spin_lock(&rtas_data_buf_lock); + memcpy(rtas_data_buf, vf_pe_array, RTAS_DATA_BUF_SIZE); + rc = rtas_call(ibm_allow_unfreeze, 5, 1, NULL, + addr, + BUID_HI(buid), + BUID_LO(buid), + rtas_data_buf, cur_vfs * sizeof(u16)); + spin_unlock(&rtas_data_buf_lock); + if (rc) + pr_warn("%s: Failed to allow unfreeze for PHB#%x-PE#%lx, rc=%x\n", + __func__, + pdn->phb->global_number, addr, rc); + return rc; +} + +static int pseries_call_allow_unfreeze(struct eeh_dev *edev) +{ + int cur_vfs = 0, rc = 0, vf_index, bus, devfn, vf_pe_num; + struct pci_dn *pdn, *tmp, *parent, *physfn_pdn; + u16 *vf_pe_array; + + vf_pe_array = kzalloc(RTAS_DATA_BUF_SIZE, GFP_KERNEL); + if (!vf_pe_array) + return -ENOMEM; + if (pci_num_vf(edev->physfn ? edev->physfn : edev->pdev)) { + if (edev->pdev->is_physfn) { + cur_vfs = pci_num_vf(edev->pdev); + pdn = eeh_dev_to_pdn(edev); + parent = pdn->parent; + for (vf_index = 0; vf_index < cur_vfs; vf_index++) + vf_pe_array[vf_index] = + cpu_to_be16(pdn->pe_num_map[vf_index]); + rc = pseries_send_allow_unfreeze(pdn, vf_pe_array, + cur_vfs); + pdn->last_allow_rc = rc; + for (vf_index = 0; vf_index < cur_vfs; vf_index++) { + list_for_each_entry_safe(pdn, tmp, + &parent->child_list, + list) { + bus = pci_iov_virtfn_bus(edev->pdev, + vf_index); + devfn = pci_iov_virtfn_devfn(edev->pdev, + vf_index); + if (pdn->busno != bus || + pdn->devfn != devfn) + continue; + pdn->last_allow_rc = rc; + } + } + } else { + pdn = pci_get_pdn(edev->pdev); + physfn_pdn = pci_get_pdn(edev->physfn); + + vf_pe_num = physfn_pdn->pe_num_map[edev->vf_index]; + vf_pe_array[0] = cpu_to_be16(vf_pe_num); + rc = pseries_send_allow_unfreeze(physfn_pdn, + vf_pe_array, 1); + pdn->last_allow_rc = rc; + } + } + + kfree(vf_pe_array); + return rc; +} + +static int pseries_notify_resume(struct eeh_dev *edev) +{ + if (!edev) + return -EEXIST; + + if (rtas_function_token(RTAS_FN_IBM_OPEN_SRIOV_ALLOW_UNFREEZE) == RTAS_UNKNOWN_SERVICE) + return -EINVAL; + + if (edev->pdev->is_physfn || edev->pdev->is_virtfn) + return pseries_call_allow_unfreeze(edev); + + return 0; +} +#endif + +static struct eeh_ops pseries_eeh_ops = { + .name = "pseries", + .probe = pseries_eeh_probe, + .set_option = pseries_eeh_set_option, + .get_state = pseries_eeh_get_state, + .reset = pseries_eeh_reset, + .get_log = pseries_eeh_get_log, + .configure_bridge = pseries_eeh_configure_bridge, + .err_inject = NULL, + .read_config = pseries_eeh_read_config, + .write_config = pseries_eeh_write_config, + .next_error = NULL, + .restore_config = NULL, /* NB: configure_bridge() does this */ +#ifdef CONFIG_PCI_IOV + .notify_resume = pseries_notify_resume +#endif +}; + +/** + * eeh_pseries_init - Register platform dependent EEH operations + * + * EEH initialization on pseries platform. This function should be + * called before any EEH related functions. + */ +static int __init eeh_pseries_init(void) +{ + struct pci_controller *phb; + struct pci_dn *pdn; + int ret, config_addr; + + /* figure out EEH RTAS function call tokens */ + ibm_set_eeh_option = rtas_function_token(RTAS_FN_IBM_SET_EEH_OPTION); + ibm_set_slot_reset = rtas_function_token(RTAS_FN_IBM_SET_SLOT_RESET); + ibm_read_slot_reset_state2 = rtas_function_token(RTAS_FN_IBM_READ_SLOT_RESET_STATE2); + ibm_read_slot_reset_state = rtas_function_token(RTAS_FN_IBM_READ_SLOT_RESET_STATE); + ibm_slot_error_detail = rtas_function_token(RTAS_FN_IBM_SLOT_ERROR_DETAIL); + ibm_get_config_addr_info2 = rtas_function_token(RTAS_FN_IBM_GET_CONFIG_ADDR_INFO2); + ibm_get_config_addr_info = rtas_function_token(RTAS_FN_IBM_GET_CONFIG_ADDR_INFO); + ibm_configure_pe = rtas_function_token(RTAS_FN_IBM_CONFIGURE_PE); + + /* + * ibm,configure-pe and ibm,configure-bridge have the same semantics, + * however ibm,configure-pe can be faster. If we can't find + * ibm,configure-pe then fall back to using ibm,configure-bridge. + */ + if (ibm_configure_pe == RTAS_UNKNOWN_SERVICE) + ibm_configure_pe = rtas_function_token(RTAS_FN_IBM_CONFIGURE_BRIDGE); + + /* + * Necessary sanity check. We needn't check "get-config-addr-info" + * and its variant since the old firmware probably support address + * of domain/bus/slot/function for EEH RTAS operations. + */ + if (ibm_set_eeh_option == RTAS_UNKNOWN_SERVICE || + ibm_set_slot_reset == RTAS_UNKNOWN_SERVICE || + (ibm_read_slot_reset_state2 == RTAS_UNKNOWN_SERVICE && + ibm_read_slot_reset_state == RTAS_UNKNOWN_SERVICE) || + ibm_slot_error_detail == RTAS_UNKNOWN_SERVICE || + ibm_configure_pe == RTAS_UNKNOWN_SERVICE) { + pr_info("EEH functionality not supported\n"); + return -EINVAL; + } + + /* Initialize error log size */ + eeh_error_buf_size = rtas_get_error_log_max(); + + /* Set EEH probe mode */ + eeh_add_flag(EEH_PROBE_MODE_DEVTREE | EEH_ENABLE_IO_FOR_LOG); + + /* Set EEH machine dependent code */ + ppc_md.pcibios_bus_add_device = pseries_pcibios_bus_add_device; + + if (is_kdump_kernel() || reset_devices) { + pr_info("Issue PHB reset ...\n"); + list_for_each_entry(phb, &hose_list, list_node) { + // Skip if the slot is empty + if (list_empty(&PCI_DN(phb->dn)->child_list)) + continue; + + pdn = list_first_entry(&PCI_DN(phb->dn)->child_list, struct pci_dn, list); + config_addr = pseries_eeh_get_pe_config_addr(pdn); + + /* invalid PE config addr */ + if (config_addr < 0) + continue; + + pseries_eeh_phb_reset(phb, config_addr, EEH_RESET_FUNDAMENTAL); + pseries_eeh_phb_reset(phb, config_addr, EEH_RESET_DEACTIVATE); + pseries_eeh_phb_configure_bridge(phb, config_addr); + } + } + + ret = eeh_init(&pseries_eeh_ops); + if (!ret) + pr_info("EEH: pSeries platform initialized\n"); + else + pr_info("EEH: pSeries platform initialization failure (%d)\n", + ret); + return ret; +} +machine_arch_initcall(pseries, eeh_pseries_init); diff --git a/arch/powerpc/platforms/pseries/event_sources.c b/arch/powerpc/platforms/pseries/event_sources.c new file mode 100644 index 000000000..623dfe0d8 --- /dev/null +++ b/arch/powerpc/platforms/pseries/event_sources.c @@ -0,0 +1,30 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2001 Dave Engebretsen IBM Corporation + */ + +#include <linux/interrupt.h> +#include <linux/of_irq.h> + +#include "pseries.h" + +void __init request_event_sources_irqs(struct device_node *np, + irq_handler_t handler, + const char *name) +{ + int i, virq, rc; + + for (i = 0; i < 16; i++) { + virq = of_irq_get(np, i); + if (virq < 0) + return; + if (WARN(!virq, "event-sources: Unable to allocate " + "interrupt number for %pOF\n", np)) + continue; + + rc = request_irq(virq, handler, 0, name, NULL); + if (WARN(rc, "event-sources: Unable to request interrupt %d for %pOF\n", + virq, np)) + return; + } +} diff --git a/arch/powerpc/platforms/pseries/firmware.c b/arch/powerpc/platforms/pseries/firmware.c new file mode 100644 index 000000000..18447e5fa --- /dev/null +++ b/arch/powerpc/platforms/pseries/firmware.c @@ -0,0 +1,191 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * pSeries firmware setup code. + * + * Portions from arch/powerpc/platforms/pseries/setup.c: + * Copyright (C) 1995 Linus Torvalds + * Adapted from 'alpha' version by Gary Thomas + * Modified by Cort Dougan (cort@cs.nmt.edu) + * Modified by PPC64 Team, IBM Corp + * + * Portions from arch/powerpc/kernel/firmware.c + * Copyright (C) 2001 Ben. Herrenschmidt (benh@kernel.crashing.org) + * Modifications for ppc64: + * Copyright (C) 2003 Dave Engebretsen <engebret@us.ibm.com> + * Copyright (C) 2005 Stephen Rothwell, IBM Corporation + * + * Copyright 2006 IBM Corporation. + */ + + +#include <linux/of_fdt.h> +#include <asm/firmware.h> +#include <asm/prom.h> +#include <asm/udbg.h> +#include <asm/svm.h> + +#include "pseries.h" + +struct hypertas_fw_feature { + unsigned long val; + char * name; +}; + +/* + * The names in this table match names in rtas/ibm,hypertas-functions. If the + * entry ends in a '*', only upto the '*' is matched. Otherwise the entire + * string must match. + */ +static __initdata struct hypertas_fw_feature +hypertas_fw_features_table[] = { + {FW_FEATURE_PFT, "hcall-pft"}, + {FW_FEATURE_TCE, "hcall-tce"}, + {FW_FEATURE_SPRG0, "hcall-sprg0"}, + {FW_FEATURE_DABR, "hcall-dabr"}, + {FW_FEATURE_COPY, "hcall-copy"}, + {FW_FEATURE_ASR, "hcall-asr"}, + {FW_FEATURE_DEBUG, "hcall-debug"}, + {FW_FEATURE_PERF, "hcall-perf"}, + {FW_FEATURE_DUMP, "hcall-dump"}, + {FW_FEATURE_INTERRUPT, "hcall-interrupt"}, + {FW_FEATURE_MIGRATE, "hcall-migrate"}, + {FW_FEATURE_PERFMON, "hcall-perfmon"}, + {FW_FEATURE_CRQ, "hcall-crq"}, + {FW_FEATURE_VIO, "hcall-vio"}, + {FW_FEATURE_RDMA, "hcall-rdma"}, + {FW_FEATURE_LLAN, "hcall-lLAN"}, + {FW_FEATURE_BULK_REMOVE, "hcall-bulk"}, + {FW_FEATURE_XDABR, "hcall-xdabr"}, + {FW_FEATURE_PUT_TCE_IND | FW_FEATURE_STUFF_TCE, + "hcall-multi-tce"}, + {FW_FEATURE_SPLPAR, "hcall-splpar"}, + {FW_FEATURE_VPHN, "hcall-vphn"}, + {FW_FEATURE_SET_MODE, "hcall-set-mode"}, + {FW_FEATURE_BEST_ENERGY, "hcall-best-energy-1*"}, + {FW_FEATURE_HPT_RESIZE, "hcall-hpt-resize"}, + {FW_FEATURE_BLOCK_REMOVE, "hcall-block-remove"}, + {FW_FEATURE_PAPR_SCM, "hcall-scm"}, + {FW_FEATURE_RPT_INVALIDATE, "hcall-rpt-invalidate"}, + {FW_FEATURE_ENERGY_SCALE_INFO, "hcall-energy-scale-info"}, + {FW_FEATURE_WATCHDOG, "hcall-watchdog"}, + {FW_FEATURE_PLPKS, "hcall-pks"}, +}; + +/* Build up the firmware features bitmask using the contents of + * device-tree/ibm,hypertas-functions. Ultimately this functionality may + * be moved into prom.c prom_init(). + */ +static void __init fw_hypertas_feature_init(const char *hypertas, + unsigned long len) +{ + const char *s; + int i; + + pr_debug(" -> fw_hypertas_feature_init()\n"); + + for (s = hypertas; s < hypertas + len; s += strlen(s) + 1) { + for (i = 0; i < ARRAY_SIZE(hypertas_fw_features_table); i++) { + const char *name = hypertas_fw_features_table[i].name; + size_t size; + + /* + * If there is a '*' at the end of name, only check + * upto there + */ + size = strlen(name); + if (size && name[size - 1] == '*') { + if (strncmp(name, s, size - 1)) + continue; + } else if (strcmp(name, s)) + continue; + + /* we have a match */ + powerpc_firmware_features |= + hypertas_fw_features_table[i].val; + break; + } + } + + if (is_secure_guest() && + (powerpc_firmware_features & FW_FEATURE_PUT_TCE_IND)) { + powerpc_firmware_features &= ~FW_FEATURE_PUT_TCE_IND; + pr_debug("SVM: disabling PUT_TCE_IND firmware feature\n"); + } + + pr_debug(" <- fw_hypertas_feature_init()\n"); +} + +struct vec5_fw_feature { + unsigned long val; + unsigned int feature; +}; + +static __initdata struct vec5_fw_feature +vec5_fw_features_table[] = { + {FW_FEATURE_FORM1_AFFINITY, OV5_FORM1_AFFINITY}, + {FW_FEATURE_PRRN, OV5_PRRN}, + {FW_FEATURE_DRMEM_V2, OV5_DRMEM_V2}, + {FW_FEATURE_DRC_INFO, OV5_DRC_INFO}, + {FW_FEATURE_FORM2_AFFINITY, OV5_FORM2_AFFINITY}, +}; + +static void __init fw_vec5_feature_init(const char *vec5, unsigned long len) +{ + unsigned int index, feat; + int i; + + pr_debug(" -> fw_vec5_feature_init()\n"); + + for (i = 0; i < ARRAY_SIZE(vec5_fw_features_table); i++) { + index = OV5_INDX(vec5_fw_features_table[i].feature); + feat = OV5_FEAT(vec5_fw_features_table[i].feature); + + if (index < len && (vec5[index] & feat)) + powerpc_firmware_features |= + vec5_fw_features_table[i].val; + } + + pr_debug(" <- fw_vec5_feature_init()\n"); +} + +/* + * Called very early, MMU is off, device-tree isn't unflattened + */ +static int __init probe_fw_features(unsigned long node, const char *uname, int + depth, void *data) +{ + const char *prop; + int len; + static int hypertas_found; + static int vec5_found; + + if (depth != 1) + return 0; + + if (!strcmp(uname, "rtas") || !strcmp(uname, "rtas@0")) { + prop = of_get_flat_dt_prop(node, "ibm,hypertas-functions", + &len); + if (prop) { + powerpc_firmware_features |= FW_FEATURE_LPAR; + fw_hypertas_feature_init(prop, len); + } + + hypertas_found = 1; + } + + if (!strcmp(uname, "chosen")) { + prop = of_get_flat_dt_prop(node, "ibm,architecture-vec-5", + &len); + if (prop) + fw_vec5_feature_init(prop, len); + + vec5_found = 1; + } + + return hypertas_found && vec5_found; +} + +void __init pseries_probe_fw_features(void) +{ + of_scan_flat_dt(probe_fw_features, NULL); +} diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c new file mode 100644 index 000000000..e62835a12 --- /dev/null +++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c @@ -0,0 +1,901 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * pseries CPU Hotplug infrastructure. + * + * Split out from arch/powerpc/platforms/pseries/setup.c + * arch/powerpc/kernel/rtas.c, and arch/powerpc/platforms/pseries/smp.c + * + * Peter Bergner, IBM March 2001. + * Copyright (C) 2001 IBM. + * Dave Engebretsen, Peter Bergner, and + * Mike Corrigan {engebret|bergner|mikec}@us.ibm.com + * Plus various changes from other IBM teams... + * + * Copyright (C) 2006 Michael Ellerman, IBM Corporation + */ + +#define pr_fmt(fmt) "pseries-hotplug-cpu: " fmt + +#include <linux/kernel.h> +#include <linux/interrupt.h> +#include <linux/delay.h> +#include <linux/sched.h> /* for idle_task_exit */ +#include <linux/sched/hotplug.h> +#include <linux/cpu.h> +#include <linux/of.h> +#include <linux/slab.h> +#include <asm/prom.h> +#include <asm/rtas.h> +#include <asm/firmware.h> +#include <asm/machdep.h> +#include <asm/vdso_datapage.h> +#include <asm/xics.h> +#include <asm/xive.h> +#include <asm/plpar_wrappers.h> +#include <asm/topology.h> + +#include "pseries.h" + +/* This version can't take the spinlock, because it never returns */ +static int rtas_stop_self_token = RTAS_UNKNOWN_SERVICE; + +/* + * Record the CPU ids used on each nodes. + * Protected by cpu_add_remove_lock. + */ +static cpumask_var_t node_recorded_ids_map[MAX_NUMNODES]; + +static void rtas_stop_self(void) +{ + static struct rtas_args args; + + local_irq_disable(); + + BUG_ON(rtas_stop_self_token == RTAS_UNKNOWN_SERVICE); + + rtas_call_unlocked(&args, rtas_stop_self_token, 0, 1, NULL); + + panic("Alas, I survived.\n"); +} + +static void pseries_cpu_offline_self(void) +{ + unsigned int hwcpu = hard_smp_processor_id(); + + local_irq_disable(); + idle_task_exit(); + if (xive_enabled()) + xive_teardown_cpu(); + else + xics_teardown_cpu(); + + unregister_slb_shadow(hwcpu); + unregister_vpa(hwcpu); + rtas_stop_self(); + + /* Should never get here... */ + BUG(); + for(;;); +} + +static int pseries_cpu_disable(void) +{ + int cpu = smp_processor_id(); + + set_cpu_online(cpu, false); + vdso_data->processorCount--; + + /*fix boot_cpuid here*/ + if (cpu == boot_cpuid) + boot_cpuid = cpumask_any(cpu_online_mask); + + /* FIXME: abstract this to not be platform specific later on */ + if (xive_enabled()) + xive_smp_disable_cpu(); + else + xics_migrate_irqs_away(); + + cleanup_cpu_mmu_context(); + + return 0; +} + +/* + * pseries_cpu_die: Wait for the cpu to die. + * @cpu: logical processor id of the CPU whose death we're awaiting. + * + * This function is called from the context of the thread which is performing + * the cpu-offline. Here we wait for long enough to allow the cpu in question + * to self-destroy so that the cpu-offline thread can send the CPU_DEAD + * notifications. + * + * OTOH, pseries_cpu_offline_self() is called by the @cpu when it wants to + * self-destruct. + */ +static void pseries_cpu_die(unsigned int cpu) +{ + int cpu_status = 1; + unsigned int pcpu = get_hard_smp_processor_id(cpu); + unsigned long timeout = jiffies + msecs_to_jiffies(120000); + + while (true) { + cpu_status = smp_query_cpu_stopped(pcpu); + if (cpu_status == QCSS_STOPPED || + cpu_status == QCSS_HARDWARE_ERROR) + break; + + if (time_after(jiffies, timeout)) { + pr_warn("CPU %i (hwid %i) didn't die after 120 seconds\n", + cpu, pcpu); + timeout = jiffies + msecs_to_jiffies(120000); + } + + cond_resched(); + } + + if (cpu_status == QCSS_HARDWARE_ERROR) { + pr_warn("CPU %i (hwid %i) reported error while dying\n", + cpu, pcpu); + } + + paca_ptrs[cpu]->cpu_start = 0; +} + +/** + * find_cpu_id_range - found a linear ranger of @nthreads free CPU ids. + * @nthreads : the number of threads (cpu ids) + * @assigned_node : the node it belongs to or NUMA_NO_NODE if free ids from any + * node can be peek. + * @cpu_mask: the returned CPU mask. + * + * Returns 0 on success. + */ +static int find_cpu_id_range(unsigned int nthreads, int assigned_node, + cpumask_var_t *cpu_mask) +{ + cpumask_var_t candidate_mask; + unsigned int cpu, node; + int rc = -ENOSPC; + + if (!zalloc_cpumask_var(&candidate_mask, GFP_KERNEL)) + return -ENOMEM; + + cpumask_clear(*cpu_mask); + for (cpu = 0; cpu < nthreads; cpu++) + cpumask_set_cpu(cpu, *cpu_mask); + + BUG_ON(!cpumask_subset(cpu_present_mask, cpu_possible_mask)); + + /* Get a bitmap of unoccupied slots. */ + cpumask_xor(candidate_mask, cpu_possible_mask, cpu_present_mask); + + if (assigned_node != NUMA_NO_NODE) { + /* + * Remove free ids previously assigned on the other nodes. We + * can walk only online nodes because once a node became online + * it is not turned offlined back. + */ + for_each_online_node(node) { + if (node == assigned_node) + continue; + cpumask_andnot(candidate_mask, candidate_mask, + node_recorded_ids_map[node]); + } + } + + if (cpumask_empty(candidate_mask)) + goto out; + + while (!cpumask_empty(*cpu_mask)) { + if (cpumask_subset(*cpu_mask, candidate_mask)) + /* Found a range where we can insert the new cpu(s) */ + break; + cpumask_shift_left(*cpu_mask, *cpu_mask, nthreads); + } + + if (!cpumask_empty(*cpu_mask)) + rc = 0; + +out: + free_cpumask_var(candidate_mask); + return rc; +} + +/* + * Update cpu_present_mask and paca(s) for a new cpu node. The wrinkle + * here is that a cpu device node may represent multiple logical cpus + * in the SMT case. We must honor the assumption in other code that + * the logical ids for sibling SMT threads x and y are adjacent, such + * that x^1 == y and y^1 == x. + */ +static int pseries_add_processor(struct device_node *np) +{ + int len, nthreads, node, cpu, assigned_node; + int rc = 0; + cpumask_var_t cpu_mask; + const __be32 *intserv; + + intserv = of_get_property(np, "ibm,ppc-interrupt-server#s", &len); + if (!intserv) + return 0; + + nthreads = len / sizeof(u32); + + if (!alloc_cpumask_var(&cpu_mask, GFP_KERNEL)) + return -ENOMEM; + + /* + * Fetch from the DT nodes read by dlpar_configure_connector() the NUMA + * node id the added CPU belongs to. + */ + node = of_node_to_nid(np); + if (node < 0 || !node_possible(node)) + node = first_online_node; + + BUG_ON(node == NUMA_NO_NODE); + assigned_node = node; + + cpu_maps_update_begin(); + + rc = find_cpu_id_range(nthreads, node, &cpu_mask); + if (rc && nr_node_ids > 1) { + /* + * Try again, considering the free CPU ids from the other node. + */ + node = NUMA_NO_NODE; + rc = find_cpu_id_range(nthreads, NUMA_NO_NODE, &cpu_mask); + } + + if (rc) { + pr_err("Cannot add cpu %pOF; this system configuration" + " supports %d logical cpus.\n", np, num_possible_cpus()); + goto out; + } + + for_each_cpu(cpu, cpu_mask) { + BUG_ON(cpu_present(cpu)); + set_cpu_present(cpu, true); + set_hard_smp_processor_id(cpu, be32_to_cpu(*intserv++)); + } + + /* Record the newly used CPU ids for the associate node. */ + cpumask_or(node_recorded_ids_map[assigned_node], + node_recorded_ids_map[assigned_node], cpu_mask); + + /* + * If node is set to NUMA_NO_NODE, CPU ids have be reused from + * another node, remove them from its mask. + */ + if (node == NUMA_NO_NODE) { + cpu = cpumask_first(cpu_mask); + pr_warn("Reusing free CPU ids %d-%d from another node\n", + cpu, cpu + nthreads - 1); + for_each_online_node(node) { + if (node == assigned_node) + continue; + cpumask_andnot(node_recorded_ids_map[node], + node_recorded_ids_map[node], + cpu_mask); + } + } + +out: + cpu_maps_update_done(); + free_cpumask_var(cpu_mask); + return rc; +} + +/* + * Update the present map for a cpu node which is going away, and set + * the hard id in the paca(s) to -1 to be consistent with boot time + * convention for non-present cpus. + */ +static void pseries_remove_processor(struct device_node *np) +{ + unsigned int cpu; + int len, nthreads, i; + const __be32 *intserv; + u32 thread; + + intserv = of_get_property(np, "ibm,ppc-interrupt-server#s", &len); + if (!intserv) + return; + + nthreads = len / sizeof(u32); + + cpu_maps_update_begin(); + for (i = 0; i < nthreads; i++) { + thread = be32_to_cpu(intserv[i]); + for_each_present_cpu(cpu) { + if (get_hard_smp_processor_id(cpu) != thread) + continue; + BUG_ON(cpu_online(cpu)); + set_cpu_present(cpu, false); + set_hard_smp_processor_id(cpu, -1); + update_numa_cpu_lookup_table(cpu, -1); + break; + } + if (cpu >= nr_cpu_ids) + printk(KERN_WARNING "Could not find cpu to remove " + "with physical id 0x%x\n", thread); + } + cpu_maps_update_done(); +} + +static int dlpar_offline_cpu(struct device_node *dn) +{ + int rc = 0; + unsigned int cpu; + int len, nthreads, i; + const __be32 *intserv; + u32 thread; + + intserv = of_get_property(dn, "ibm,ppc-interrupt-server#s", &len); + if (!intserv) + return -EINVAL; + + nthreads = len / sizeof(u32); + + cpu_maps_update_begin(); + for (i = 0; i < nthreads; i++) { + thread = be32_to_cpu(intserv[i]); + for_each_present_cpu(cpu) { + if (get_hard_smp_processor_id(cpu) != thread) + continue; + + if (!cpu_online(cpu)) + break; + + /* + * device_offline() will return -EBUSY (via cpu_down()) if there + * is only one CPU left. Check it here to fail earlier and with a + * more informative error message, while also retaining the + * cpu_add_remove_lock to be sure that no CPUs are being + * online/offlined during this check. + */ + if (num_online_cpus() == 1) { + pr_warn("Unable to remove last online CPU %pOFn\n", dn); + rc = -EBUSY; + goto out_unlock; + } + + cpu_maps_update_done(); + rc = device_offline(get_cpu_device(cpu)); + if (rc) + goto out; + cpu_maps_update_begin(); + break; + } + if (cpu == num_possible_cpus()) { + pr_warn("Could not find cpu to offline with physical id 0x%x\n", + thread); + } + } +out_unlock: + cpu_maps_update_done(); + +out: + return rc; +} + +static int dlpar_online_cpu(struct device_node *dn) +{ + int rc = 0; + unsigned int cpu; + int len, nthreads, i; + const __be32 *intserv; + u32 thread; + + intserv = of_get_property(dn, "ibm,ppc-interrupt-server#s", &len); + if (!intserv) + return -EINVAL; + + nthreads = len / sizeof(u32); + + cpu_maps_update_begin(); + for (i = 0; i < nthreads; i++) { + thread = be32_to_cpu(intserv[i]); + for_each_present_cpu(cpu) { + if (get_hard_smp_processor_id(cpu) != thread) + continue; + + if (!topology_is_primary_thread(cpu)) { + if (cpu_smt_control != CPU_SMT_ENABLED) + break; + if (!topology_smt_thread_allowed(cpu)) + break; + } + + cpu_maps_update_done(); + find_and_update_cpu_nid(cpu); + rc = device_online(get_cpu_device(cpu)); + if (rc) { + dlpar_offline_cpu(dn); + goto out; + } + cpu_maps_update_begin(); + + break; + } + if (cpu == num_possible_cpus()) + printk(KERN_WARNING "Could not find cpu to online " + "with physical id 0x%x\n", thread); + } + cpu_maps_update_done(); + +out: + return rc; + +} + +static bool dlpar_cpu_exists(struct device_node *parent, u32 drc_index) +{ + struct device_node *child = NULL; + u32 my_drc_index; + bool found; + int rc; + + /* Assume cpu doesn't exist */ + found = false; + + for_each_child_of_node(parent, child) { + rc = of_property_read_u32(child, "ibm,my-drc-index", + &my_drc_index); + if (rc) + continue; + + if (my_drc_index == drc_index) { + of_node_put(child); + found = true; + break; + } + } + + return found; +} + +static bool drc_info_valid_index(struct device_node *parent, u32 drc_index) +{ + struct property *info; + struct of_drc_info drc; + const __be32 *value; + u32 index; + int count, i, j; + + info = of_find_property(parent, "ibm,drc-info", NULL); + if (!info) + return false; + + value = of_prop_next_u32(info, NULL, &count); + + /* First value of ibm,drc-info is number of drc-info records */ + if (value) + value++; + else + return false; + + for (i = 0; i < count; i++) { + if (of_read_drc_info_cell(&info, &value, &drc)) + return false; + + if (strncmp(drc.drc_type, "CPU", 3)) + break; + + if (drc_index > drc.last_drc_index) + continue; + + index = drc.drc_index_start; + for (j = 0; j < drc.num_sequential_elems; j++) { + if (drc_index == index) + return true; + + index += drc.sequential_inc; + } + } + + return false; +} + +static bool valid_cpu_drc_index(struct device_node *parent, u32 drc_index) +{ + bool found = false; + int rc, index; + + if (of_property_present(parent, "ibm,drc-info")) + return drc_info_valid_index(parent, drc_index); + + /* Note that the format of the ibm,drc-indexes array is + * the number of entries in the array followed by the array + * of drc values so we start looking at index = 1. + */ + index = 1; + while (!found) { + u32 drc; + + rc = of_property_read_u32_index(parent, "ibm,drc-indexes", + index++, &drc); + + if (rc) + break; + + if (drc == drc_index) + found = true; + } + + return found; +} + +static int pseries_cpuhp_attach_nodes(struct device_node *dn) +{ + struct of_changeset cs; + int ret; + + /* + * This device node is unattached but may have siblings; open-code the + * traversal. + */ + for (of_changeset_init(&cs); dn != NULL; dn = dn->sibling) { + ret = of_changeset_attach_node(&cs, dn); + if (ret) + goto out; + } + + ret = of_changeset_apply(&cs); +out: + of_changeset_destroy(&cs); + return ret; +} + +static ssize_t dlpar_cpu_add(u32 drc_index) +{ + struct device_node *dn, *parent; + int rc, saved_rc; + + pr_debug("Attempting to add CPU, drc index: %x\n", drc_index); + + parent = of_find_node_by_path("/cpus"); + if (!parent) { + pr_warn("Failed to find CPU root node \"/cpus\"\n"); + return -ENODEV; + } + + if (dlpar_cpu_exists(parent, drc_index)) { + of_node_put(parent); + pr_warn("CPU with drc index %x already exists\n", drc_index); + return -EINVAL; + } + + if (!valid_cpu_drc_index(parent, drc_index)) { + of_node_put(parent); + pr_warn("Cannot find CPU (drc index %x) to add.\n", drc_index); + return -EINVAL; + } + + rc = dlpar_acquire_drc(drc_index); + if (rc) { + pr_warn("Failed to acquire DRC, rc: %d, drc index: %x\n", + rc, drc_index); + of_node_put(parent); + return -EINVAL; + } + + dn = dlpar_configure_connector(cpu_to_be32(drc_index), parent); + if (!dn) { + pr_warn("Failed call to configure-connector, drc index: %x\n", + drc_index); + dlpar_release_drc(drc_index); + of_node_put(parent); + return -EINVAL; + } + + rc = pseries_cpuhp_attach_nodes(dn); + + /* Regardless we are done with parent now */ + of_node_put(parent); + + if (rc) { + saved_rc = rc; + pr_warn("Failed to attach node %pOFn, rc: %d, drc index: %x\n", + dn, rc, drc_index); + + rc = dlpar_release_drc(drc_index); + if (!rc) + dlpar_free_cc_nodes(dn); + + return saved_rc; + } + + update_numa_distance(dn); + + rc = dlpar_online_cpu(dn); + if (rc) { + saved_rc = rc; + pr_warn("Failed to online cpu %pOFn, rc: %d, drc index: %x\n", + dn, rc, drc_index); + + rc = dlpar_detach_node(dn); + if (!rc) + dlpar_release_drc(drc_index); + + return saved_rc; + } + + pr_debug("Successfully added CPU %pOFn, drc index: %x\n", dn, + drc_index); + return rc; +} + +static unsigned int pseries_cpuhp_cache_use_count(const struct device_node *cachedn) +{ + unsigned int use_count = 0; + struct device_node *dn, *tn; + + WARN_ON(!of_node_is_type(cachedn, "cache")); + + for_each_of_cpu_node(dn) { + tn = of_find_next_cache_node(dn); + of_node_put(tn); + if (tn == cachedn) + use_count++; + } + + for_each_node_by_type(dn, "cache") { + tn = of_find_next_cache_node(dn); + of_node_put(tn); + if (tn == cachedn) + use_count++; + } + + return use_count; +} + +static int pseries_cpuhp_detach_nodes(struct device_node *cpudn) +{ + struct device_node *dn; + struct of_changeset cs; + int ret = 0; + + of_changeset_init(&cs); + ret = of_changeset_detach_node(&cs, cpudn); + if (ret) + goto out; + + dn = cpudn; + while ((dn = of_find_next_cache_node(dn))) { + if (pseries_cpuhp_cache_use_count(dn) > 1) { + of_node_put(dn); + break; + } + + ret = of_changeset_detach_node(&cs, dn); + of_node_put(dn); + if (ret) + goto out; + } + + ret = of_changeset_apply(&cs); +out: + of_changeset_destroy(&cs); + return ret; +} + +static ssize_t dlpar_cpu_remove(struct device_node *dn, u32 drc_index) +{ + int rc; + + pr_debug("Attempting to remove CPU %pOFn, drc index: %x\n", + dn, drc_index); + + rc = dlpar_offline_cpu(dn); + if (rc) { + pr_warn("Failed to offline CPU %pOFn, rc: %d\n", dn, rc); + return -EINVAL; + } + + rc = dlpar_release_drc(drc_index); + if (rc) { + pr_warn("Failed to release drc (%x) for CPU %pOFn, rc: %d\n", + drc_index, dn, rc); + dlpar_online_cpu(dn); + return rc; + } + + rc = pseries_cpuhp_detach_nodes(dn); + if (rc) { + int saved_rc = rc; + + pr_warn("Failed to detach CPU %pOFn, rc: %d", dn, rc); + + rc = dlpar_acquire_drc(drc_index); + if (!rc) + dlpar_online_cpu(dn); + + return saved_rc; + } + + pr_debug("Successfully removed CPU, drc index: %x\n", drc_index); + return 0; +} + +static struct device_node *cpu_drc_index_to_dn(u32 drc_index) +{ + struct device_node *dn; + u32 my_index; + int rc; + + for_each_node_by_type(dn, "cpu") { + rc = of_property_read_u32(dn, "ibm,my-drc-index", &my_index); + if (rc) + continue; + + if (my_index == drc_index) + break; + } + + return dn; +} + +static int dlpar_cpu_remove_by_index(u32 drc_index) +{ + struct device_node *dn; + int rc; + + dn = cpu_drc_index_to_dn(drc_index); + if (!dn) { + pr_warn("Cannot find CPU (drc index %x) to remove\n", + drc_index); + return -ENODEV; + } + + rc = dlpar_cpu_remove(dn, drc_index); + of_node_put(dn); + return rc; +} + +int dlpar_cpu(struct pseries_hp_errorlog *hp_elog) +{ + u32 drc_index; + int rc; + + drc_index = hp_elog->_drc_u.drc_index; + + lock_device_hotplug(); + + switch (hp_elog->action) { + case PSERIES_HP_ELOG_ACTION_REMOVE: + if (hp_elog->id_type == PSERIES_HP_ELOG_ID_DRC_INDEX) { + rc = dlpar_cpu_remove_by_index(drc_index); + /* + * Setting the isolation state of an UNISOLATED/CONFIGURED + * device to UNISOLATE is a no-op, but the hypervisor can + * use it as a hint that the CPU removal failed. + */ + if (rc) + dlpar_unisolate_drc(drc_index); + } + else + rc = -EINVAL; + break; + case PSERIES_HP_ELOG_ACTION_ADD: + if (hp_elog->id_type == PSERIES_HP_ELOG_ID_DRC_INDEX) + rc = dlpar_cpu_add(drc_index); + else + rc = -EINVAL; + break; + default: + pr_err("Invalid action (%d) specified\n", hp_elog->action); + rc = -EINVAL; + break; + } + + unlock_device_hotplug(); + return rc; +} + +#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE + +static ssize_t dlpar_cpu_probe(const char *buf, size_t count) +{ + u32 drc_index; + int rc; + + rc = kstrtou32(buf, 0, &drc_index); + if (rc) + return -EINVAL; + + rc = dlpar_cpu_add(drc_index); + + return rc ? rc : count; +} + +static ssize_t dlpar_cpu_release(const char *buf, size_t count) +{ + struct device_node *dn; + u32 drc_index; + int rc; + + dn = of_find_node_by_path(buf); + if (!dn) + return -EINVAL; + + rc = of_property_read_u32(dn, "ibm,my-drc-index", &drc_index); + if (rc) { + of_node_put(dn); + return -EINVAL; + } + + rc = dlpar_cpu_remove(dn, drc_index); + of_node_put(dn); + + return rc ? rc : count; +} + +#endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */ + +static int pseries_smp_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct of_reconfig_data *rd = data; + int err = 0; + + switch (action) { + case OF_RECONFIG_ATTACH_NODE: + err = pseries_add_processor(rd->dn); + break; + case OF_RECONFIG_DETACH_NODE: + pseries_remove_processor(rd->dn); + break; + } + return notifier_from_errno(err); +} + +static struct notifier_block pseries_smp_nb = { + .notifier_call = pseries_smp_notifier, +}; + +void __init pseries_cpu_hotplug_init(void) +{ + int qcss_tok; + + rtas_stop_self_token = rtas_function_token(RTAS_FN_STOP_SELF); + qcss_tok = rtas_function_token(RTAS_FN_QUERY_CPU_STOPPED_STATE); + + if (rtas_stop_self_token == RTAS_UNKNOWN_SERVICE || + qcss_tok == RTAS_UNKNOWN_SERVICE) { + printk(KERN_INFO "CPU Hotplug not supported by firmware " + "- disabling.\n"); + return; + } + + smp_ops->cpu_offline_self = pseries_cpu_offline_self; + smp_ops->cpu_disable = pseries_cpu_disable; + smp_ops->cpu_die = pseries_cpu_die; +} + +static int __init pseries_dlpar_init(void) +{ + unsigned int node; + +#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE + ppc_md.cpu_probe = dlpar_cpu_probe; + ppc_md.cpu_release = dlpar_cpu_release; +#endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */ + + /* Processors can be added/removed only on LPAR */ + if (firmware_has_feature(FW_FEATURE_LPAR)) { + for_each_node(node) { + if (!alloc_cpumask_var_node(&node_recorded_ids_map[node], + GFP_KERNEL, node)) + return -ENOMEM; + + /* Record ids of CPU added at boot time */ + cpumask_copy(node_recorded_ids_map[node], + cpumask_of_node(node)); + } + + of_reconfig_notifier_register(&pseries_smp_nb); + } + + return 0; +} +machine_arch_initcall(pseries, pseries_dlpar_init); diff --git a/arch/powerpc/platforms/pseries/hotplug-memory.c b/arch/powerpc/platforms/pseries/hotplug-memory.c new file mode 100644 index 000000000..4adca5b61 --- /dev/null +++ b/arch/powerpc/platforms/pseries/hotplug-memory.c @@ -0,0 +1,923 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * pseries Memory Hotplug infrastructure. + * + * Copyright (C) 2008 Badari Pulavarty, IBM Corporation + */ + +#define pr_fmt(fmt) "pseries-hotplug-mem: " fmt + +#include <linux/of.h> +#include <linux/of_address.h> +#include <linux/memblock.h> +#include <linux/memory.h> +#include <linux/memory_hotplug.h> +#include <linux/slab.h> + +#include <asm/firmware.h> +#include <asm/machdep.h> +#include <asm/sparsemem.h> +#include <asm/fadump.h> +#include <asm/drmem.h> +#include "pseries.h" + +static void dlpar_free_property(struct property *prop) +{ + kfree(prop->name); + kfree(prop->value); + kfree(prop); +} + +static struct property *dlpar_clone_property(struct property *prop, + u32 prop_size) +{ + struct property *new_prop; + + new_prop = kzalloc(sizeof(*new_prop), GFP_KERNEL); + if (!new_prop) + return NULL; + + new_prop->name = kstrdup(prop->name, GFP_KERNEL); + new_prop->value = kzalloc(prop_size, GFP_KERNEL); + if (!new_prop->name || !new_prop->value) { + dlpar_free_property(new_prop); + return NULL; + } + + memcpy(new_prop->value, prop->value, prop->length); + new_prop->length = prop_size; + + of_property_set_flag(new_prop, OF_DYNAMIC); + return new_prop; +} + +static bool find_aa_index(struct device_node *dr_node, + struct property *ala_prop, + const u32 *lmb_assoc, u32 *aa_index) +{ + u32 *assoc_arrays, new_prop_size; + struct property *new_prop; + int aa_arrays, aa_array_entries, aa_array_sz; + int i, index; + + /* + * The ibm,associativity-lookup-arrays property is defined to be + * a 32-bit value specifying the number of associativity arrays + * followed by a 32-bitvalue specifying the number of entries per + * array, followed by the associativity arrays. + */ + assoc_arrays = ala_prop->value; + + aa_arrays = be32_to_cpu(assoc_arrays[0]); + aa_array_entries = be32_to_cpu(assoc_arrays[1]); + aa_array_sz = aa_array_entries * sizeof(u32); + + for (i = 0; i < aa_arrays; i++) { + index = (i * aa_array_entries) + 2; + + if (memcmp(&assoc_arrays[index], &lmb_assoc[1], aa_array_sz)) + continue; + + *aa_index = i; + return true; + } + + new_prop_size = ala_prop->length + aa_array_sz; + new_prop = dlpar_clone_property(ala_prop, new_prop_size); + if (!new_prop) + return false; + + assoc_arrays = new_prop->value; + + /* increment the number of entries in the lookup array */ + assoc_arrays[0] = cpu_to_be32(aa_arrays + 1); + + /* copy the new associativity into the lookup array */ + index = aa_arrays * aa_array_entries + 2; + memcpy(&assoc_arrays[index], &lmb_assoc[1], aa_array_sz); + + of_update_property(dr_node, new_prop); + + /* + * The associativity lookup array index for this lmb is + * number of entries - 1 since we added its associativity + * to the end of the lookup array. + */ + *aa_index = be32_to_cpu(assoc_arrays[0]) - 1; + return true; +} + +static int update_lmb_associativity_index(struct drmem_lmb *lmb) +{ + struct device_node *parent, *lmb_node, *dr_node; + struct property *ala_prop; + const u32 *lmb_assoc; + u32 aa_index; + bool found; + + parent = of_find_node_by_path("/"); + if (!parent) + return -ENODEV; + + lmb_node = dlpar_configure_connector(cpu_to_be32(lmb->drc_index), + parent); + of_node_put(parent); + if (!lmb_node) + return -EINVAL; + + lmb_assoc = of_get_property(lmb_node, "ibm,associativity", NULL); + if (!lmb_assoc) { + dlpar_free_cc_nodes(lmb_node); + return -ENODEV; + } + + update_numa_distance(lmb_node); + + dr_node = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory"); + if (!dr_node) { + dlpar_free_cc_nodes(lmb_node); + return -ENODEV; + } + + ala_prop = of_find_property(dr_node, "ibm,associativity-lookup-arrays", + NULL); + if (!ala_prop) { + of_node_put(dr_node); + dlpar_free_cc_nodes(lmb_node); + return -ENODEV; + } + + found = find_aa_index(dr_node, ala_prop, lmb_assoc, &aa_index); + + of_node_put(dr_node); + dlpar_free_cc_nodes(lmb_node); + + if (!found) { + pr_err("Could not find LMB associativity\n"); + return -1; + } + + lmb->aa_index = aa_index; + return 0; +} + +static struct memory_block *lmb_to_memblock(struct drmem_lmb *lmb) +{ + unsigned long section_nr; + struct memory_block *mem_block; + + section_nr = pfn_to_section_nr(PFN_DOWN(lmb->base_addr)); + + mem_block = find_memory_block(section_nr); + return mem_block; +} + +static int get_lmb_range(u32 drc_index, int n_lmbs, + struct drmem_lmb **start_lmb, + struct drmem_lmb **end_lmb) +{ + struct drmem_lmb *lmb, *start, *end; + struct drmem_lmb *limit; + + start = NULL; + for_each_drmem_lmb(lmb) { + if (lmb->drc_index == drc_index) { + start = lmb; + break; + } + } + + if (!start) + return -EINVAL; + + end = &start[n_lmbs]; + + limit = &drmem_info->lmbs[drmem_info->n_lmbs]; + if (end > limit) + return -EINVAL; + + *start_lmb = start; + *end_lmb = end; + return 0; +} + +static int dlpar_change_lmb_state(struct drmem_lmb *lmb, bool online) +{ + struct memory_block *mem_block; + int rc; + + mem_block = lmb_to_memblock(lmb); + if (!mem_block) + return -EINVAL; + + if (online && mem_block->dev.offline) + rc = device_online(&mem_block->dev); + else if (!online && !mem_block->dev.offline) + rc = device_offline(&mem_block->dev); + else + rc = 0; + + put_device(&mem_block->dev); + + return rc; +} + +static int dlpar_online_lmb(struct drmem_lmb *lmb) +{ + return dlpar_change_lmb_state(lmb, true); +} + +#ifdef CONFIG_MEMORY_HOTREMOVE +static int dlpar_offline_lmb(struct drmem_lmb *lmb) +{ + return dlpar_change_lmb_state(lmb, false); +} + +static int pseries_remove_memblock(unsigned long base, unsigned long memblock_size) +{ + unsigned long start_pfn; + int sections_per_block; + int i; + + start_pfn = base >> PAGE_SHIFT; + + lock_device_hotplug(); + + if (!pfn_valid(start_pfn)) + goto out; + + sections_per_block = memory_block_size / MIN_MEMORY_BLOCK_SIZE; + + for (i = 0; i < sections_per_block; i++) { + __remove_memory(base, MIN_MEMORY_BLOCK_SIZE); + base += MIN_MEMORY_BLOCK_SIZE; + } + +out: + /* Update memory regions for memory remove */ + memblock_remove(base, memblock_size); + unlock_device_hotplug(); + return 0; +} + +static int pseries_remove_mem_node(struct device_node *np) +{ + int ret; + struct resource res; + + /* + * Check to see if we are actually removing memory + */ + if (!of_node_is_type(np, "memory")) + return 0; + + /* + * Find the base address and size of the memblock + */ + ret = of_address_to_resource(np, 0, &res); + if (ret) + return ret; + + pseries_remove_memblock(res.start, resource_size(&res)); + return 0; +} + +static bool lmb_is_removable(struct drmem_lmb *lmb) +{ + if ((lmb->flags & DRCONF_MEM_RESERVED) || + !(lmb->flags & DRCONF_MEM_ASSIGNED)) + return false; + +#ifdef CONFIG_FA_DUMP + /* + * Don't hot-remove memory that falls in fadump boot memory area + * and memory that is reserved for capturing old kernel memory. + */ + if (is_fadump_memory_area(lmb->base_addr, memory_block_size_bytes())) + return false; +#endif + /* device_offline() will determine if we can actually remove this lmb */ + return true; +} + +static int dlpar_add_lmb(struct drmem_lmb *); + +static int dlpar_remove_lmb(struct drmem_lmb *lmb) +{ + struct memory_block *mem_block; + int rc; + + if (!lmb_is_removable(lmb)) + return -EINVAL; + + mem_block = lmb_to_memblock(lmb); + if (mem_block == NULL) + return -EINVAL; + + rc = dlpar_offline_lmb(lmb); + if (rc) { + put_device(&mem_block->dev); + return rc; + } + + __remove_memory(lmb->base_addr, memory_block_size); + put_device(&mem_block->dev); + + /* Update memory regions for memory remove */ + memblock_remove(lmb->base_addr, memory_block_size); + + invalidate_lmb_associativity_index(lmb); + lmb->flags &= ~DRCONF_MEM_ASSIGNED; + + return 0; +} + +static int dlpar_memory_remove_by_count(u32 lmbs_to_remove) +{ + struct drmem_lmb *lmb; + int lmbs_reserved = 0; + int lmbs_available = 0; + int rc; + + pr_info("Attempting to hot-remove %d LMB(s)\n", lmbs_to_remove); + + if (lmbs_to_remove == 0) + return -EINVAL; + + /* Validate that there are enough LMBs to satisfy the request */ + for_each_drmem_lmb(lmb) { + if (lmb_is_removable(lmb)) + lmbs_available++; + + if (lmbs_available == lmbs_to_remove) + break; + } + + if (lmbs_available < lmbs_to_remove) { + pr_info("Not enough LMBs available (%d of %d) to satisfy request\n", + lmbs_available, lmbs_to_remove); + return -EINVAL; + } + + for_each_drmem_lmb(lmb) { + rc = dlpar_remove_lmb(lmb); + if (rc) + continue; + + /* Mark this lmb so we can add it later if all of the + * requested LMBs cannot be removed. + */ + drmem_mark_lmb_reserved(lmb); + + lmbs_reserved++; + if (lmbs_reserved == lmbs_to_remove) + break; + } + + if (lmbs_reserved != lmbs_to_remove) { + pr_err("Memory hot-remove failed, adding LMB's back\n"); + + for_each_drmem_lmb(lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + rc = dlpar_add_lmb(lmb); + if (rc) + pr_err("Failed to add LMB back, drc index %x\n", + lmb->drc_index); + + drmem_remove_lmb_reservation(lmb); + + lmbs_reserved--; + if (lmbs_reserved == 0) + break; + } + + rc = -EINVAL; + } else { + for_each_drmem_lmb(lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + dlpar_release_drc(lmb->drc_index); + pr_info("Memory at %llx was hot-removed\n", + lmb->base_addr); + + drmem_remove_lmb_reservation(lmb); + + lmbs_reserved--; + if (lmbs_reserved == 0) + break; + } + rc = 0; + } + + return rc; +} + +static int dlpar_memory_remove_by_index(u32 drc_index) +{ + struct drmem_lmb *lmb; + int lmb_found; + int rc; + + pr_debug("Attempting to hot-remove LMB, drc index %x\n", drc_index); + + lmb_found = 0; + for_each_drmem_lmb(lmb) { + if (lmb->drc_index == drc_index) { + lmb_found = 1; + rc = dlpar_remove_lmb(lmb); + if (!rc) + dlpar_release_drc(lmb->drc_index); + + break; + } + } + + if (!lmb_found) { + pr_debug("Failed to look up LMB for drc index %x\n", drc_index); + rc = -EINVAL; + } else if (rc) { + pr_debug("Failed to hot-remove memory at %llx\n", + lmb->base_addr); + } else { + pr_debug("Memory at %llx was hot-removed\n", lmb->base_addr); + } + + return rc; +} + +static int dlpar_memory_remove_by_ic(u32 lmbs_to_remove, u32 drc_index) +{ + struct drmem_lmb *lmb, *start_lmb, *end_lmb; + int rc; + + pr_info("Attempting to hot-remove %u LMB(s) at %x\n", + lmbs_to_remove, drc_index); + + if (lmbs_to_remove == 0) + return -EINVAL; + + rc = get_lmb_range(drc_index, lmbs_to_remove, &start_lmb, &end_lmb); + if (rc) + return -EINVAL; + + /* + * Validate that all LMBs in range are not reserved. Note that it + * is ok if they are !ASSIGNED since our goal here is to remove the + * LMB range, regardless of whether some LMBs were already removed + * by any other reason. + * + * This is a contrast to what is done in remove_by_count() where we + * check for both RESERVED and !ASSIGNED (via lmb_is_removable()), + * because we want to remove a fixed amount of LMBs in that function. + */ + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + if (lmb->flags & DRCONF_MEM_RESERVED) { + pr_err("Memory at %llx (drc index %x) is reserved\n", + lmb->base_addr, lmb->drc_index); + return -EINVAL; + } + } + + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + /* + * dlpar_remove_lmb() will error out if the LMB is already + * !ASSIGNED, but this case is a no-op for us. + */ + if (!(lmb->flags & DRCONF_MEM_ASSIGNED)) + continue; + + rc = dlpar_remove_lmb(lmb); + if (rc) + break; + + drmem_mark_lmb_reserved(lmb); + } + + if (rc) { + pr_err("Memory indexed-count-remove failed, adding any removed LMBs\n"); + + + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + /* + * Setting the isolation state of an UNISOLATED/CONFIGURED + * device to UNISOLATE is a no-op, but the hypervisor can + * use it as a hint that the LMB removal failed. + */ + dlpar_unisolate_drc(lmb->drc_index); + + rc = dlpar_add_lmb(lmb); + if (rc) + pr_err("Failed to add LMB, drc index %x\n", + lmb->drc_index); + + drmem_remove_lmb_reservation(lmb); + } + rc = -EINVAL; + } else { + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + dlpar_release_drc(lmb->drc_index); + pr_info("Memory at %llx (drc index %x) was hot-removed\n", + lmb->base_addr, lmb->drc_index); + + drmem_remove_lmb_reservation(lmb); + } + } + + return rc; +} + +#else +static inline int pseries_remove_memblock(unsigned long base, + unsigned long memblock_size) +{ + return -EOPNOTSUPP; +} +static inline int pseries_remove_mem_node(struct device_node *np) +{ + return 0; +} +static int dlpar_remove_lmb(struct drmem_lmb *lmb) +{ + return -EOPNOTSUPP; +} +static int dlpar_memory_remove_by_count(u32 lmbs_to_remove) +{ + return -EOPNOTSUPP; +} +static int dlpar_memory_remove_by_index(u32 drc_index) +{ + return -EOPNOTSUPP; +} + +static int dlpar_memory_remove_by_ic(u32 lmbs_to_remove, u32 drc_index) +{ + return -EOPNOTSUPP; +} +#endif /* CONFIG_MEMORY_HOTREMOVE */ + +static int dlpar_add_lmb(struct drmem_lmb *lmb) +{ + unsigned long block_sz; + int nid, rc; + + if (lmb->flags & DRCONF_MEM_ASSIGNED) + return -EINVAL; + + rc = update_lmb_associativity_index(lmb); + if (rc) { + dlpar_release_drc(lmb->drc_index); + return rc; + } + + block_sz = memory_block_size_bytes(); + + /* Find the node id for this LMB. Fake one if necessary. */ + nid = of_drconf_to_nid_single(lmb); + if (nid < 0 || !node_possible(nid)) + nid = first_online_node; + + /* Add the memory */ + rc = __add_memory(nid, lmb->base_addr, block_sz, MHP_MEMMAP_ON_MEMORY); + if (rc) { + invalidate_lmb_associativity_index(lmb); + return rc; + } + + rc = dlpar_online_lmb(lmb); + if (rc) { + __remove_memory(lmb->base_addr, block_sz); + invalidate_lmb_associativity_index(lmb); + } else { + lmb->flags |= DRCONF_MEM_ASSIGNED; + } + + return rc; +} + +static int dlpar_memory_add_by_count(u32 lmbs_to_add) +{ + struct drmem_lmb *lmb; + int lmbs_available = 0; + int lmbs_reserved = 0; + int rc; + + pr_info("Attempting to hot-add %d LMB(s)\n", lmbs_to_add); + + if (lmbs_to_add == 0) + return -EINVAL; + + /* Validate that there are enough LMBs to satisfy the request */ + for_each_drmem_lmb(lmb) { + if (lmb->flags & DRCONF_MEM_RESERVED) + continue; + + if (!(lmb->flags & DRCONF_MEM_ASSIGNED)) + lmbs_available++; + + if (lmbs_available == lmbs_to_add) + break; + } + + if (lmbs_available < lmbs_to_add) + return -EINVAL; + + for_each_drmem_lmb(lmb) { + if (lmb->flags & DRCONF_MEM_ASSIGNED) + continue; + + rc = dlpar_acquire_drc(lmb->drc_index); + if (rc) + continue; + + rc = dlpar_add_lmb(lmb); + if (rc) { + dlpar_release_drc(lmb->drc_index); + continue; + } + + /* Mark this lmb so we can remove it later if all of the + * requested LMBs cannot be added. + */ + drmem_mark_lmb_reserved(lmb); + lmbs_reserved++; + if (lmbs_reserved == lmbs_to_add) + break; + } + + if (lmbs_reserved != lmbs_to_add) { + pr_err("Memory hot-add failed, removing any added LMBs\n"); + + for_each_drmem_lmb(lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + rc = dlpar_remove_lmb(lmb); + if (rc) + pr_err("Failed to remove LMB, drc index %x\n", + lmb->drc_index); + else + dlpar_release_drc(lmb->drc_index); + + drmem_remove_lmb_reservation(lmb); + lmbs_reserved--; + + if (lmbs_reserved == 0) + break; + } + rc = -EINVAL; + } else { + for_each_drmem_lmb(lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + pr_debug("Memory at %llx (drc index %x) was hot-added\n", + lmb->base_addr, lmb->drc_index); + drmem_remove_lmb_reservation(lmb); + lmbs_reserved--; + + if (lmbs_reserved == 0) + break; + } + rc = 0; + } + + return rc; +} + +static int dlpar_memory_add_by_index(u32 drc_index) +{ + struct drmem_lmb *lmb; + int rc, lmb_found; + + pr_info("Attempting to hot-add LMB, drc index %x\n", drc_index); + + lmb_found = 0; + for_each_drmem_lmb(lmb) { + if (lmb->drc_index == drc_index) { + lmb_found = 1; + rc = dlpar_acquire_drc(lmb->drc_index); + if (!rc) { + rc = dlpar_add_lmb(lmb); + if (rc) + dlpar_release_drc(lmb->drc_index); + } + + break; + } + } + + if (!lmb_found) + rc = -EINVAL; + + if (rc) + pr_info("Failed to hot-add memory, drc index %x\n", drc_index); + else + pr_info("Memory at %llx (drc index %x) was hot-added\n", + lmb->base_addr, drc_index); + + return rc; +} + +static int dlpar_memory_add_by_ic(u32 lmbs_to_add, u32 drc_index) +{ + struct drmem_lmb *lmb, *start_lmb, *end_lmb; + int rc; + + pr_info("Attempting to hot-add %u LMB(s) at index %x\n", + lmbs_to_add, drc_index); + + if (lmbs_to_add == 0) + return -EINVAL; + + rc = get_lmb_range(drc_index, lmbs_to_add, &start_lmb, &end_lmb); + if (rc) + return -EINVAL; + + /* Validate that the LMBs in this range are not reserved */ + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + /* Fail immediately if the whole range can't be hot-added */ + if (lmb->flags & DRCONF_MEM_RESERVED) { + pr_err("Memory at %llx (drc index %x) is reserved\n", + lmb->base_addr, lmb->drc_index); + return -EINVAL; + } + } + + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + if (lmb->flags & DRCONF_MEM_ASSIGNED) + continue; + + rc = dlpar_acquire_drc(lmb->drc_index); + if (rc) + break; + + rc = dlpar_add_lmb(lmb); + if (rc) { + dlpar_release_drc(lmb->drc_index); + break; + } + + drmem_mark_lmb_reserved(lmb); + } + + if (rc) { + pr_err("Memory indexed-count-add failed, removing any added LMBs\n"); + + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + rc = dlpar_remove_lmb(lmb); + if (rc) + pr_err("Failed to remove LMB, drc index %x\n", + lmb->drc_index); + else + dlpar_release_drc(lmb->drc_index); + + drmem_remove_lmb_reservation(lmb); + } + rc = -EINVAL; + } else { + for_each_drmem_lmb_in_range(lmb, start_lmb, end_lmb) { + if (!drmem_lmb_reserved(lmb)) + continue; + + pr_info("Memory at %llx (drc index %x) was hot-added\n", + lmb->base_addr, lmb->drc_index); + drmem_remove_lmb_reservation(lmb); + } + } + + return rc; +} + +int dlpar_memory(struct pseries_hp_errorlog *hp_elog) +{ + u32 count, drc_index; + int rc; + + lock_device_hotplug(); + + switch (hp_elog->action) { + case PSERIES_HP_ELOG_ACTION_ADD: + switch (hp_elog->id_type) { + case PSERIES_HP_ELOG_ID_DRC_COUNT: + count = hp_elog->_drc_u.drc_count; + rc = dlpar_memory_add_by_count(count); + break; + case PSERIES_HP_ELOG_ID_DRC_INDEX: + drc_index = hp_elog->_drc_u.drc_index; + rc = dlpar_memory_add_by_index(drc_index); + break; + case PSERIES_HP_ELOG_ID_DRC_IC: + count = hp_elog->_drc_u.ic.count; + drc_index = hp_elog->_drc_u.ic.index; + rc = dlpar_memory_add_by_ic(count, drc_index); + break; + default: + rc = -EINVAL; + break; + } + + break; + case PSERIES_HP_ELOG_ACTION_REMOVE: + switch (hp_elog->id_type) { + case PSERIES_HP_ELOG_ID_DRC_COUNT: + count = hp_elog->_drc_u.drc_count; + rc = dlpar_memory_remove_by_count(count); + break; + case PSERIES_HP_ELOG_ID_DRC_INDEX: + drc_index = hp_elog->_drc_u.drc_index; + rc = dlpar_memory_remove_by_index(drc_index); + break; + case PSERIES_HP_ELOG_ID_DRC_IC: + count = hp_elog->_drc_u.ic.count; + drc_index = hp_elog->_drc_u.ic.index; + rc = dlpar_memory_remove_by_ic(count, drc_index); + break; + default: + rc = -EINVAL; + break; + } + + break; + default: + pr_err("Invalid action (%d) specified\n", hp_elog->action); + rc = -EINVAL; + break; + } + + if (!rc) + rc = drmem_update_dt(); + + unlock_device_hotplug(); + return rc; +} + +static int pseries_add_mem_node(struct device_node *np) +{ + int ret; + struct resource res; + + /* + * Check to see if we are actually adding memory + */ + if (!of_node_is_type(np, "memory")) + return 0; + + /* + * Find the base and size of the memblock + */ + ret = of_address_to_resource(np, 0, &res); + if (ret) + return ret; + + /* + * Update memory region to represent the memory add + */ + ret = memblock_add(res.start, resource_size(&res)); + return (ret < 0) ? -EINVAL : 0; +} + +static int pseries_memory_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct of_reconfig_data *rd = data; + int err = 0; + + switch (action) { + case OF_RECONFIG_ATTACH_NODE: + err = pseries_add_mem_node(rd->dn); + break; + case OF_RECONFIG_DETACH_NODE: + err = pseries_remove_mem_node(rd->dn); + break; + case OF_RECONFIG_UPDATE_PROPERTY: + if (!strcmp(rd->dn->name, + "ibm,dynamic-reconfiguration-memory")) + drmem_update_lmbs(rd->prop); + } + return notifier_from_errno(err); +} + +static struct notifier_block pseries_mem_nb = { + .notifier_call = pseries_memory_notifier, +}; + +static int __init pseries_memory_hotplug_init(void) +{ + if (firmware_has_feature(FW_FEATURE_LPAR)) + of_reconfig_notifier_register(&pseries_mem_nb); + + return 0; +} +machine_device_initcall(pseries, pseries_memory_hotplug_init); diff --git a/arch/powerpc/platforms/pseries/hvCall.S b/arch/powerpc/platforms/pseries/hvCall.S new file mode 100644 index 000000000..2b0cac6fb --- /dev/null +++ b/arch/powerpc/platforms/pseries/hvCall.S @@ -0,0 +1,370 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * This file contains the generic code to perform a call to the + * pSeries LPAR hypervisor. + */ +#include <linux/jump_label.h> +#include <asm/hvcall.h> +#include <asm/processor.h> +#include <asm/ppc_asm.h> +#include <asm/asm-offsets.h> +#include <asm/ptrace.h> +#include <asm/feature-fixups.h> + + .section ".text" + +#ifdef CONFIG_TRACEPOINTS + +#ifndef CONFIG_JUMP_LABEL + .data + + .globl hcall_tracepoint_refcount +hcall_tracepoint_refcount: + .8byte 0 + + .section ".text" +#endif + +/* + * precall must preserve all registers. use unused STK_PARAM() + * areas to save snapshots and opcode. STK_PARAM() in the caller's + * frame will be available even on ELFv2 because these are all + * variadic functions. + */ +#define HCALL_INST_PRECALL(FIRST_REG) \ + mflr r0; \ + std r3,STK_PARAM(R3)(r1); \ + std r4,STK_PARAM(R4)(r1); \ + std r5,STK_PARAM(R5)(r1); \ + std r6,STK_PARAM(R6)(r1); \ + std r7,STK_PARAM(R7)(r1); \ + std r8,STK_PARAM(R8)(r1); \ + std r9,STK_PARAM(R9)(r1); \ + std r10,STK_PARAM(R10)(r1); \ + std r0,16(r1); \ + addi r4,r1,STK_PARAM(FIRST_REG); \ + stdu r1,-STACK_FRAME_MIN_SIZE(r1); \ + bl CFUNC(__trace_hcall_entry); \ + ld r3,STACK_FRAME_MIN_SIZE+STK_PARAM(R3)(r1); \ + ld r4,STACK_FRAME_MIN_SIZE+STK_PARAM(R4)(r1); \ + ld r5,STACK_FRAME_MIN_SIZE+STK_PARAM(R5)(r1); \ + ld r6,STACK_FRAME_MIN_SIZE+STK_PARAM(R6)(r1); \ + ld r7,STACK_FRAME_MIN_SIZE+STK_PARAM(R7)(r1); \ + ld r8,STACK_FRAME_MIN_SIZE+STK_PARAM(R8)(r1); \ + ld r9,STACK_FRAME_MIN_SIZE+STK_PARAM(R9)(r1); \ + ld r10,STACK_FRAME_MIN_SIZE+STK_PARAM(R10)(r1) + +/* + * postcall is performed immediately before function return which + * allows liberal use of volatile registers. + */ +#define __HCALL_INST_POSTCALL \ + ld r0,STACK_FRAME_MIN_SIZE+STK_PARAM(R3)(r1); \ + std r3,STACK_FRAME_MIN_SIZE+STK_PARAM(R3)(r1); \ + mr r4,r3; \ + mr r3,r0; \ + bl CFUNC(__trace_hcall_exit); \ + ld r0,STACK_FRAME_MIN_SIZE+16(r1); \ + addi r1,r1,STACK_FRAME_MIN_SIZE; \ + ld r3,STK_PARAM(R3)(r1); \ + mtlr r0 + +#define HCALL_INST_POSTCALL_NORETS \ + li r5,0; \ + __HCALL_INST_POSTCALL + +#define HCALL_INST_POSTCALL(BUFREG) \ + mr r5,BUFREG; \ + __HCALL_INST_POSTCALL + +#ifdef CONFIG_JUMP_LABEL +#define HCALL_BRANCH(LABEL) \ + ARCH_STATIC_BRANCH(LABEL, hcall_tracepoint_key) +#else + +/* + * We branch around this in early init (eg when populating the MMU + * hashtable) by using an unconditional cpu feature. + */ +#define HCALL_BRANCH(LABEL) \ +BEGIN_FTR_SECTION; \ + b 1f; \ +END_FTR_SECTION(0, 1); \ + LOAD_REG_ADDR(r12, hcall_tracepoint_refcount) ; \ + ld r12,0(r12); \ + cmpdi r12,0; \ + bne- LABEL; \ +1: +#endif + +#else +#define HCALL_INST_PRECALL(FIRST_ARG) +#define HCALL_INST_POSTCALL_NORETS +#define HCALL_INST_POSTCALL(BUFREG) +#define HCALL_BRANCH(LABEL) +#endif + +_GLOBAL_TOC(plpar_hcall_norets_notrace) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + HVSC /* invoke the hypervisor */ + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + blr /* return r3 = status */ + +_GLOBAL_TOC(plpar_hcall_norets) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + HCALL_BRANCH(plpar_hcall_norets_trace) + HVSC /* invoke the hypervisor */ + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + blr /* return r3 = status */ + +#ifdef CONFIG_TRACEPOINTS +plpar_hcall_norets_trace: + HCALL_INST_PRECALL(R4) + HVSC + HCALL_INST_POSTCALL_NORETS + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + blr +#endif + +_GLOBAL_TOC(plpar_hcall) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_BRANCH(plpar_hcall_trace) + + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + + HVSC /* invoke the hypervisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +#ifdef CONFIG_TRACEPOINTS +plpar_hcall_trace: + HCALL_INST_PRECALL(R5) + + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + + HVSC + + ld r12,STACK_FRAME_MIN_SIZE+STK_PARAM(R4)(r1) + std r4,0(r12) + std r5,8(r12) + std r6,16(r12) + std r7,24(r12) + + HCALL_INST_POSTCALL(r12) + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr +#endif + +/* + * plpar_hcall_raw can be called in real mode. kexec/kdump need some + * hypervisor calls to be executed in real mode. So plpar_hcall_raw + * does not access the per cpu hypervisor call statistics variables, + * since these variables may not be present in the RMO region. + */ +_GLOBAL(plpar_hcall_raw) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + + HVSC /* invoke the hypervisor */ + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +_GLOBAL_TOC(plpar_hcall9) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_BRANCH(plpar_hcall9_trace) + + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + ld r10,STK_PARAM(R11)(r1) /* put arg7 in R10 */ + ld r11,STK_PARAM(R12)(r1) /* put arg8 in R11 */ + ld r12,STK_PARAM(R13)(r1) /* put arg9 in R12 */ + + HVSC /* invoke the hypervisor */ + + mr r0,r12 + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + std r8, 32(r12) + std r9, 40(r12) + std r10,48(r12) + std r11,56(r12) + std r0, 64(r12) + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +#ifdef CONFIG_TRACEPOINTS +plpar_hcall9_trace: + HCALL_INST_PRECALL(R5) + + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + ld r10,STACK_FRAME_MIN_SIZE+STK_PARAM(R11)(r1) + ld r11,STACK_FRAME_MIN_SIZE+STK_PARAM(R12)(r1) + ld r12,STACK_FRAME_MIN_SIZE+STK_PARAM(R13)(r1) + + HVSC + + mr r0,r12 + ld r12,STACK_FRAME_MIN_SIZE+STK_PARAM(R4)(r1) + std r4,0(r12) + std r5,8(r12) + std r6,16(r12) + std r7,24(r12) + std r8,32(r12) + std r9,40(r12) + std r10,48(r12) + std r11,56(r12) + std r0,64(r12) + + HCALL_INST_POSTCALL(r12) + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr +#endif + +/* See plpar_hcall_raw to see why this is needed */ +_GLOBAL(plpar_hcall9_raw) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + std r4,STK_PARAM(R4)(r1) /* Save ret buffer */ + + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + mr r9,r10 + ld r10,STK_PARAM(R11)(r1) /* put arg7 in R10 */ + ld r11,STK_PARAM(R12)(r1) /* put arg8 in R11 */ + ld r12,STK_PARAM(R13)(r1) /* put arg9 in R12 */ + + HVSC /* invoke the hypervisor */ + + mr r0,r12 + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + std r8, 32(r12) + std r9, 40(r12) + std r10,48(r12) + std r11,56(r12) + std r0, 64(r12) + + li r4,0 + stb r4,PACASRR_VALID(r13) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ diff --git a/arch/powerpc/platforms/pseries/hvCall_inst.c b/arch/powerpc/platforms/pseries/hvCall_inst.c new file mode 100644 index 000000000..3a50612a7 --- /dev/null +++ b/arch/powerpc/platforms/pseries/hvCall_inst.c @@ -0,0 +1,140 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2006 Mike Kravetz IBM Corporation + * + * Hypervisor Call Instrumentation + */ + +#include <linux/kernel.h> +#include <linux/percpu.h> +#include <linux/debugfs.h> +#include <linux/seq_file.h> +#include <linux/cpumask.h> +#include <asm/hvcall.h> +#include <asm/firmware.h> +#include <asm/cputable.h> +#include <asm/trace.h> +#include <asm/machdep.h> + +/* For hcall instrumentation. One structure per-hcall, per-CPU */ +struct hcall_stats { + unsigned long num_calls; /* number of calls (on this CPU) */ + unsigned long tb_total; /* total wall time (mftb) of calls. */ + unsigned long purr_total; /* total cpu time (PURR) of calls. */ + unsigned long tb_start; + unsigned long purr_start; +}; +#define HCALL_STAT_ARRAY_SIZE ((MAX_HCALL_OPCODE >> 2) + 1) + +static DEFINE_PER_CPU(struct hcall_stats[HCALL_STAT_ARRAY_SIZE], hcall_stats); + +/* + * Routines for displaying the statistics in debugfs + */ +static void *hc_start(struct seq_file *m, loff_t *pos) +{ + if ((int)*pos < (HCALL_STAT_ARRAY_SIZE-1)) + return (void *)(unsigned long)(*pos + 1); + + return NULL; +} + +static void *hc_next(struct seq_file *m, void *p, loff_t * pos) +{ + ++*pos; + + return hc_start(m, pos); +} + +static void hc_stop(struct seq_file *m, void *p) +{ +} + +static int hc_show(struct seq_file *m, void *p) +{ + unsigned long h_num = (unsigned long)p; + struct hcall_stats *hs = m->private; + + if (hs[h_num].num_calls) { + if (cpu_has_feature(CPU_FTR_PURR)) + seq_printf(m, "%lu %lu %lu %lu\n", h_num<<2, + hs[h_num].num_calls, + hs[h_num].tb_total, + hs[h_num].purr_total); + else + seq_printf(m, "%lu %lu %lu\n", h_num<<2, + hs[h_num].num_calls, + hs[h_num].tb_total); + } + + return 0; +} + +static const struct seq_operations hcall_inst_sops = { + .start = hc_start, + .next = hc_next, + .stop = hc_stop, + .show = hc_show +}; + +DEFINE_SEQ_ATTRIBUTE(hcall_inst); + +#define HCALL_ROOT_DIR "hcall_inst" +#define CPU_NAME_BUF_SIZE 32 + + +static void probe_hcall_entry(void *ignored, unsigned long opcode, unsigned long *args) +{ + struct hcall_stats *h; + + if (opcode > MAX_HCALL_OPCODE) + return; + + h = this_cpu_ptr(&hcall_stats[opcode / 4]); + h->tb_start = mftb(); + h->purr_start = mfspr(SPRN_PURR); +} + +static void probe_hcall_exit(void *ignored, unsigned long opcode, long retval, + unsigned long *retbuf) +{ + struct hcall_stats *h; + + if (opcode > MAX_HCALL_OPCODE) + return; + + h = this_cpu_ptr(&hcall_stats[opcode / 4]); + h->num_calls++; + h->tb_total += mftb() - h->tb_start; + h->purr_total += mfspr(SPRN_PURR) - h->purr_start; +} + +static int __init hcall_inst_init(void) +{ + struct dentry *hcall_root; + char cpu_name_buf[CPU_NAME_BUF_SIZE]; + int cpu; + + if (!firmware_has_feature(FW_FEATURE_LPAR)) + return 0; + + if (register_trace_hcall_entry(probe_hcall_entry, NULL)) + return -EINVAL; + + if (register_trace_hcall_exit(probe_hcall_exit, NULL)) { + unregister_trace_hcall_entry(probe_hcall_entry, NULL); + return -EINVAL; + } + + hcall_root = debugfs_create_dir(HCALL_ROOT_DIR, NULL); + + for_each_possible_cpu(cpu) { + snprintf(cpu_name_buf, CPU_NAME_BUF_SIZE, "cpu%d", cpu); + debugfs_create_file(cpu_name_buf, 0444, hcall_root, + per_cpu(hcall_stats, cpu), + &hcall_inst_fops); + } + + return 0; +} +machine_device_initcall(pseries, hcall_inst_init); diff --git a/arch/powerpc/platforms/pseries/hvconsole.c b/arch/powerpc/platforms/pseries/hvconsole.c new file mode 100644 index 000000000..1ac52963e --- /dev/null +++ b/arch/powerpc/platforms/pseries/hvconsole.c @@ -0,0 +1,75 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * hvconsole.c + * Copyright (C) 2004 Hollis Blanchard, IBM Corporation + * Copyright (C) 2004 IBM Corporation + * + * Additional Author(s): + * Ryan S. Arnold <rsa@us.ibm.com> + * + * LPAR console support. + */ + +#include <linux/kernel.h> +#include <linux/export.h> +#include <linux/errno.h> +#include <asm/hvcall.h> +#include <asm/hvconsole.h> +#include <asm/plpar_wrappers.h> + +/** + * hvc_get_chars - retrieve characters from firmware for denoted vterm adapter + * @vtermno: The vtermno or unit_address of the adapter from which to fetch the + * data. + * @buf: The character buffer into which to put the character data fetched from + * firmware. + * @count: not used? + */ +int hvc_get_chars(uint32_t vtermno, char *buf, int count) +{ + long ret; + unsigned long retbuf[PLPAR_HCALL_BUFSIZE]; + unsigned long *lbuf = (unsigned long *)buf; + + ret = plpar_hcall(H_GET_TERM_CHAR, retbuf, vtermno); + lbuf[0] = be64_to_cpu(retbuf[1]); + lbuf[1] = be64_to_cpu(retbuf[2]); + + if (ret == H_SUCCESS) + return retbuf[0]; + + return 0; +} + +EXPORT_SYMBOL(hvc_get_chars); + + +/** + * hvc_put_chars: send characters to firmware for denoted vterm adapter + * @vtermno: The vtermno or unit_address of the adapter from which the data + * originated. + * @buf: The character buffer that contains the character data to send to + * firmware. Must be at least 16 bytes, even if count is less than 16. + * @count: Send this number of characters. + */ +int hvc_put_chars(uint32_t vtermno, const char *buf, int count) +{ + unsigned long *lbuf = (unsigned long *) buf; + long ret; + + + /* hcall will ret H_PARAMETER if 'count' exceeds firmware max.*/ + if (count > MAX_VIO_PUT_CHARS) + count = MAX_VIO_PUT_CHARS; + + ret = plpar_hcall_norets(H_PUT_TERM_CHAR, vtermno, count, + cpu_to_be64(lbuf[0]), + cpu_to_be64(lbuf[1])); + if (ret == H_SUCCESS) + return count; + if (ret == H_BUSY) + return -EAGAIN; + return -EIO; +} + +EXPORT_SYMBOL(hvc_put_chars); diff --git a/arch/powerpc/platforms/pseries/hvcserver.c b/arch/powerpc/platforms/pseries/hvcserver.c new file mode 100644 index 000000000..d48c9c7ce --- /dev/null +++ b/arch/powerpc/platforms/pseries/hvcserver.c @@ -0,0 +1,239 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * hvcserver.c + * Copyright (C) 2004 Ryan S Arnold, IBM Corporation + * + * PPC64 virtual I/O console server support. + */ + +#include <linux/kernel.h> +#include <linux/list.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/string.h> + +#include <asm/hvcall.h> +#include <asm/hvcserver.h> +#include <asm/io.h> + +#define HVCS_ARCH_VERSION "1.0.0" + +MODULE_AUTHOR("Ryan S. Arnold <rsa@us.ibm.com>"); +MODULE_DESCRIPTION("IBM hvcs ppc64 API"); +MODULE_LICENSE("GPL"); +MODULE_VERSION(HVCS_ARCH_VERSION); + +/* + * Convert arch specific return codes into relevant errnos. The hvcs + * functions aren't performance sensitive, so this conversion isn't an + * issue. + */ +static int hvcs_convert(long to_convert) +{ + switch (to_convert) { + case H_SUCCESS: + return 0; + case H_PARAMETER: + return -EINVAL; + case H_HARDWARE: + return -EIO; + case H_BUSY: + case H_LONG_BUSY_ORDER_1_MSEC: + case H_LONG_BUSY_ORDER_10_MSEC: + case H_LONG_BUSY_ORDER_100_MSEC: + case H_LONG_BUSY_ORDER_1_SEC: + case H_LONG_BUSY_ORDER_10_SEC: + case H_LONG_BUSY_ORDER_100_SEC: + return -EBUSY; + case H_FUNCTION: + default: + return -EPERM; + } +} + +/** + * hvcs_free_partner_info - free pi allocated by hvcs_get_partner_info + * @head: list_head pointer for an allocated list of partner info structs to + * free. + * + * This function is used to free the partner info list that was returned by + * calling hvcs_get_partner_info(). + */ +int hvcs_free_partner_info(struct list_head *head) +{ + struct hvcs_partner_info *pi; + struct list_head *element; + + if (!head) + return -EINVAL; + + while (!list_empty(head)) { + element = head->next; + pi = list_entry(element, struct hvcs_partner_info, node); + list_del(element); + kfree(pi); + } + + return 0; +} +EXPORT_SYMBOL(hvcs_free_partner_info); + +/* Helper function for hvcs_get_partner_info */ +static int hvcs_next_partner(uint32_t unit_address, + unsigned long last_p_partition_ID, + unsigned long last_p_unit_address, unsigned long *pi_buff) + +{ + long retval; + retval = plpar_hcall_norets(H_VTERM_PARTNER_INFO, unit_address, + last_p_partition_ID, + last_p_unit_address, virt_to_phys(pi_buff)); + return hvcs_convert(retval); +} + +/** + * hvcs_get_partner_info - Get all of the partner info for a vty-server adapter + * @unit_address: The unit_address of the vty-server adapter for which this + * function is fetching partner info. + * @head: An initialized list_head pointer to an empty list to use to return the + * list of partner info fetched from the hypervisor to the caller. + * @pi_buff: A page sized buffer pre-allocated prior to calling this function + * that is to be used to be used by firmware as an iterator to keep track + * of the partner info retrieval. + * + * This function returns non-zero on success, or if there is no partner info. + * + * The pi_buff is pre-allocated prior to calling this function because this + * function may be called with a spin_lock held and kmalloc of a page is not + * recommended as GFP_ATOMIC. + * + * The first long of this buffer is used to store a partner unit address. The + * second long is used to store a partner partition ID and starting at + * pi_buff[2] is the 79 character Converged Location Code (diff size than the + * unsigned longs, hence the casting mumbo jumbo you see later). + * + * Invocation of this function should always be followed by an invocation of + * hvcs_free_partner_info() using a pointer to the SAME list head instance + * that was passed as a parameter to this function. + */ +int hvcs_get_partner_info(uint32_t unit_address, struct list_head *head, + unsigned long *pi_buff) +{ + /* + * Dealt with as longs because of the hcall interface even though the + * values are uint32_t. + */ + unsigned long last_p_partition_ID; + unsigned long last_p_unit_address; + struct hvcs_partner_info *next_partner_info = NULL; + int more = 1; + int retval; + + /* invalid parameters */ + if (!head || !pi_buff) + return -EINVAL; + + memset(pi_buff, 0x00, PAGE_SIZE); + last_p_partition_ID = last_p_unit_address = ~0UL; + INIT_LIST_HEAD(head); + + do { + retval = hvcs_next_partner(unit_address, last_p_partition_ID, + last_p_unit_address, pi_buff); + if (retval) { + /* + * Don't indicate that we've failed if we have + * any list elements. + */ + if (!list_empty(head)) + return 0; + return retval; + } + + last_p_partition_ID = be64_to_cpu(pi_buff[0]); + last_p_unit_address = be64_to_cpu(pi_buff[1]); + + /* This indicates that there are no further partners */ + if (last_p_partition_ID == ~0UL + && last_p_unit_address == ~0UL) + break; + + /* This is a very small struct and will be freed soon in + * hvcs_free_partner_info(). */ + next_partner_info = kmalloc(sizeof(struct hvcs_partner_info), + GFP_ATOMIC); + + if (!next_partner_info) { + printk(KERN_WARNING "HVCONSOLE: kmalloc() failed to" + " allocate partner info struct.\n"); + hvcs_free_partner_info(head); + return -ENOMEM; + } + + next_partner_info->unit_address + = (unsigned int)last_p_unit_address; + next_partner_info->partition_ID + = (unsigned int)last_p_partition_ID; + + /* copy the Null-term char too */ + strscpy(&next_partner_info->location_code[0], + (char *)&pi_buff[2], + sizeof(next_partner_info->location_code)); + + list_add_tail(&(next_partner_info->node), head); + next_partner_info = NULL; + + } while (more); + + return 0; +} +EXPORT_SYMBOL(hvcs_get_partner_info); + +/** + * hvcs_register_connection - establish a connection between this vty-server and + * a vty. + * @unit_address: The unit address of the vty-server adapter that is to be + * establish a connection. + * @p_partition_ID: The partition ID of the vty adapter that is to be connected. + * @p_unit_address: The unit address of the vty adapter to which the vty-server + * is to be connected. + * + * If this function is called once and -EINVAL is returned it may + * indicate that the partner info needs to be refreshed for the + * target unit address at which point the caller must invoke + * hvcs_get_partner_info() and then call this function again. If, + * for a second time, -EINVAL is returned then it indicates that + * there is probably already a partner connection registered to a + * different vty-server adapter. It is also possible that a second + * -EINVAL may indicate that one of the parms is not valid, for + * instance if the link was removed between the vty-server adapter + * and the vty adapter that you are trying to open. Don't shoot the + * messenger. Firmware implemented it this way. + */ +int hvcs_register_connection( uint32_t unit_address, + uint32_t p_partition_ID, uint32_t p_unit_address) +{ + long retval; + retval = plpar_hcall_norets(H_REGISTER_VTERM, unit_address, + p_partition_ID, p_unit_address); + return hvcs_convert(retval); +} +EXPORT_SYMBOL(hvcs_register_connection); + +/** + * hvcs_free_connection - free the connection between a vty-server and vty + * @unit_address: The unit address of the vty-server that is to have its + * connection severed. + * + * This function is used to free the partner connection between a vty-server + * adapter and a vty adapter. + * + * If -EBUSY is returned continue to call this function until 0 is returned. + */ +int hvcs_free_connection(uint32_t unit_address) +{ + long retval; + retval = plpar_hcall_norets(H_FREE_VTERM, unit_address); + return hvcs_convert(retval); +} +EXPORT_SYMBOL(hvcs_free_connection); diff --git a/arch/powerpc/platforms/pseries/ibmebus.c b/arch/powerpc/platforms/pseries/ibmebus.c new file mode 100644 index 000000000..998e3aff2 --- /dev/null +++ b/arch/powerpc/platforms/pseries/ibmebus.c @@ -0,0 +1,479 @@ +/* + * IBM PowerPC IBM eBus Infrastructure Support. + * + * Copyright (c) 2005 IBM Corporation + * Joachim Fenkes <fenkes@de.ibm.com> + * Heiko J Schick <schickhj@de.ibm.com> + * + * All rights reserved. + * + * This source code is distributed under a dual license of GPL v2.0 and OpenIB + * BSD. + * + * OpenIB BSD License + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are met: + * + * Redistributions of source code must retain the above copyright notice, this + * list of conditions and the following disclaimer. + * + * Redistributions in binary form must reproduce the above copyright notice, + * this list of conditions and the following disclaimer in the documentation + * and/or other materials + * provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" + * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE + * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR + * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER + * IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE + * POSSIBILITY OF SUCH DAMAGE. + */ + +#include <linux/init.h> +#include <linux/export.h> +#include <linux/console.h> +#include <linux/kobject.h> +#include <linux/dma-map-ops.h> +#include <linux/interrupt.h> +#include <linux/irqdomain.h> +#include <linux/of.h> +#include <linux/slab.h> +#include <linux/stat.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> +#include <asm/ibmebus.h> +#include <asm/machdep.h> + +static struct device ibmebus_bus_device = { /* fake "parent" device */ + .init_name = "ibmebus", +}; + +struct bus_type ibmebus_bus_type; + +/* These devices will automatically be added to the bus during init */ +static const struct of_device_id ibmebus_matches[] __initconst = { + { .compatible = "IBM,lhca" }, + { .compatible = "IBM,lhea" }, + {}, +}; + +static void *ibmebus_alloc_coherent(struct device *dev, + size_t size, + dma_addr_t *dma_handle, + gfp_t flag, + unsigned long attrs) +{ + void *mem; + + mem = kmalloc(size, flag); + *dma_handle = (dma_addr_t)mem; + + return mem; +} + +static void ibmebus_free_coherent(struct device *dev, + size_t size, void *vaddr, + dma_addr_t dma_handle, + unsigned long attrs) +{ + kfree(vaddr); +} + +static dma_addr_t ibmebus_map_page(struct device *dev, + struct page *page, + unsigned long offset, + size_t size, + enum dma_data_direction direction, + unsigned long attrs) +{ + return (dma_addr_t)(page_address(page) + offset); +} + +static void ibmebus_unmap_page(struct device *dev, + dma_addr_t dma_addr, + size_t size, + enum dma_data_direction direction, + unsigned long attrs) +{ + return; +} + +static int ibmebus_map_sg(struct device *dev, + struct scatterlist *sgl, + int nents, enum dma_data_direction direction, + unsigned long attrs) +{ + struct scatterlist *sg; + int i; + + for_each_sg(sgl, sg, nents, i) { + sg->dma_address = (dma_addr_t) sg_virt(sg); + sg->dma_length = sg->length; + } + + return nents; +} + +static void ibmebus_unmap_sg(struct device *dev, + struct scatterlist *sg, + int nents, enum dma_data_direction direction, + unsigned long attrs) +{ + return; +} + +static int ibmebus_dma_supported(struct device *dev, u64 mask) +{ + return mask == DMA_BIT_MASK(64); +} + +static u64 ibmebus_dma_get_required_mask(struct device *dev) +{ + return DMA_BIT_MASK(64); +} + +static const struct dma_map_ops ibmebus_dma_ops = { + .alloc = ibmebus_alloc_coherent, + .free = ibmebus_free_coherent, + .map_sg = ibmebus_map_sg, + .unmap_sg = ibmebus_unmap_sg, + .dma_supported = ibmebus_dma_supported, + .get_required_mask = ibmebus_dma_get_required_mask, + .map_page = ibmebus_map_page, + .unmap_page = ibmebus_unmap_page, +}; + +static int ibmebus_match_path(struct device *dev, const void *data) +{ + struct device_node *dn = to_platform_device(dev)->dev.of_node; + struct device_node *tn = of_find_node_by_path(data); + + of_node_put(tn); + + return (tn == dn); +} + +static int ibmebus_match_node(struct device *dev, const void *data) +{ + return to_platform_device(dev)->dev.of_node == data; +} + +static int ibmebus_create_device(struct device_node *dn) +{ + struct platform_device *dev; + int ret; + + dev = of_device_alloc(dn, NULL, &ibmebus_bus_device); + if (!dev) + return -ENOMEM; + + dev->dev.bus = &ibmebus_bus_type; + dev->dev.dma_ops = &ibmebus_dma_ops; + + ret = of_device_add(dev); + if (ret) + platform_device_put(dev); + return ret; +} + +static int ibmebus_create_devices(const struct of_device_id *matches) +{ + struct device_node *root, *child; + struct device *dev; + int ret = 0; + + root = of_find_node_by_path("/"); + + for_each_child_of_node(root, child) { + if (!of_match_node(matches, child)) + continue; + + dev = bus_find_device(&ibmebus_bus_type, NULL, child, + ibmebus_match_node); + if (dev) { + put_device(dev); + continue; + } + + ret = ibmebus_create_device(child); + if (ret) { + printk(KERN_ERR "%s: failed to create device (%i)", + __func__, ret); + of_node_put(child); + break; + } + } + + of_node_put(root); + return ret; +} + +int ibmebus_register_driver(struct platform_driver *drv) +{ + /* If the driver uses devices that ibmebus doesn't know, add them */ + ibmebus_create_devices(drv->driver.of_match_table); + + drv->driver.bus = &ibmebus_bus_type; + return driver_register(&drv->driver); +} +EXPORT_SYMBOL(ibmebus_register_driver); + +void ibmebus_unregister_driver(struct platform_driver *drv) +{ + driver_unregister(&drv->driver); +} +EXPORT_SYMBOL(ibmebus_unregister_driver); + +int ibmebus_request_irq(u32 ist, irq_handler_t handler, + unsigned long irq_flags, const char *devname, + void *dev_id) +{ + unsigned int irq = irq_create_mapping(NULL, ist); + + if (!irq) + return -EINVAL; + + return request_irq(irq, handler, irq_flags, devname, dev_id); +} +EXPORT_SYMBOL(ibmebus_request_irq); + +void ibmebus_free_irq(u32 ist, void *dev_id) +{ + unsigned int irq = irq_find_mapping(NULL, ist); + + free_irq(irq, dev_id); + irq_dispose_mapping(irq); +} +EXPORT_SYMBOL(ibmebus_free_irq); + +static char *ibmebus_chomp(const char *in, size_t count) +{ + char *out = kmalloc(count + 1, GFP_KERNEL); + + if (!out) + return NULL; + + memcpy(out, in, count); + out[count] = '\0'; + if (out[count - 1] == '\n') + out[count - 1] = '\0'; + + return out; +} + +static ssize_t probe_store(const struct bus_type *bus, const char *buf, size_t count) +{ + struct device_node *dn = NULL; + struct device *dev; + char *path; + ssize_t rc = 0; + + path = ibmebus_chomp(buf, count); + if (!path) + return -ENOMEM; + + dev = bus_find_device(&ibmebus_bus_type, NULL, path, + ibmebus_match_path); + if (dev) { + put_device(dev); + printk(KERN_WARNING "%s: %s has already been probed\n", + __func__, path); + rc = -EEXIST; + goto out; + } + + if ((dn = of_find_node_by_path(path))) { + rc = ibmebus_create_device(dn); + of_node_put(dn); + } else { + printk(KERN_WARNING "%s: no such device node: %s\n", + __func__, path); + rc = -ENODEV; + } + +out: + kfree(path); + if (rc) + return rc; + return count; +} +static BUS_ATTR_WO(probe); + +static ssize_t remove_store(const struct bus_type *bus, const char *buf, size_t count) +{ + struct device *dev; + char *path; + + path = ibmebus_chomp(buf, count); + if (!path) + return -ENOMEM; + + if ((dev = bus_find_device(&ibmebus_bus_type, NULL, path, + ibmebus_match_path))) { + of_device_unregister(to_platform_device(dev)); + put_device(dev); + + kfree(path); + return count; + } else { + printk(KERN_WARNING "%s: %s not on the bus\n", + __func__, path); + + kfree(path); + return -ENODEV; + } +} +static BUS_ATTR_WO(remove); + +static struct attribute *ibmbus_bus_attrs[] = { + &bus_attr_probe.attr, + &bus_attr_remove.attr, + NULL, +}; +ATTRIBUTE_GROUPS(ibmbus_bus); + +static int ibmebus_bus_bus_match(struct device *dev, struct device_driver *drv) +{ + const struct of_device_id *matches = drv->of_match_table; + + if (!matches) + return 0; + + return of_match_device(matches, dev) != NULL; +} + +static int ibmebus_bus_device_probe(struct device *dev) +{ + int error = -ENODEV; + struct platform_driver *drv; + struct platform_device *of_dev; + + drv = to_platform_driver(dev->driver); + of_dev = to_platform_device(dev); + + if (!drv->probe) + return error; + + get_device(dev); + + if (of_driver_match_device(dev, dev->driver)) + error = drv->probe(of_dev); + if (error) + put_device(dev); + + return error; +} + +static void ibmebus_bus_device_remove(struct device *dev) +{ + struct platform_device *of_dev = to_platform_device(dev); + struct platform_driver *drv = to_platform_driver(dev->driver); + + if (dev->driver && drv->remove) + drv->remove(of_dev); +} + +static void ibmebus_bus_device_shutdown(struct device *dev) +{ + struct platform_device *of_dev = to_platform_device(dev); + struct platform_driver *drv = to_platform_driver(dev->driver); + + if (dev->driver && drv->shutdown) + drv->shutdown(of_dev); +} + +/* + * ibmebus_bus_device_attrs + */ +static ssize_t devspec_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct platform_device *ofdev; + + ofdev = to_platform_device(dev); + return sprintf(buf, "%pOF\n", ofdev->dev.of_node); +} +static DEVICE_ATTR_RO(devspec); + +static ssize_t name_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct platform_device *ofdev; + + ofdev = to_platform_device(dev); + return sprintf(buf, "%pOFn\n", ofdev->dev.of_node); +} +static DEVICE_ATTR_RO(name); + +static ssize_t modalias_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return of_device_modalias(dev, buf, PAGE_SIZE); +} +static DEVICE_ATTR_RO(modalias); + +static struct attribute *ibmebus_bus_device_attrs[] = { + &dev_attr_devspec.attr, + &dev_attr_name.attr, + &dev_attr_modalias.attr, + NULL, +}; +ATTRIBUTE_GROUPS(ibmebus_bus_device); + +static int ibmebus_bus_modalias(const struct device *dev, struct kobj_uevent_env *env) +{ + return of_device_uevent_modalias(dev, env); +} + +struct bus_type ibmebus_bus_type = { + .name = "ibmebus", + .uevent = ibmebus_bus_modalias, + .bus_groups = ibmbus_bus_groups, + .match = ibmebus_bus_bus_match, + .probe = ibmebus_bus_device_probe, + .remove = ibmebus_bus_device_remove, + .shutdown = ibmebus_bus_device_shutdown, + .dev_groups = ibmebus_bus_device_groups, +}; +EXPORT_SYMBOL(ibmebus_bus_type); + +static int __init ibmebus_bus_init(void) +{ + int err; + + printk(KERN_INFO "IBM eBus Device Driver\n"); + + err = bus_register(&ibmebus_bus_type); + if (err) { + printk(KERN_ERR "%s: failed to register IBM eBus.\n", + __func__); + return err; + } + + err = device_register(&ibmebus_bus_device); + if (err) { + printk(KERN_WARNING "%s: device_register returned %i\n", + __func__, err); + put_device(&ibmebus_bus_device); + bus_unregister(&ibmebus_bus_type); + + return err; + } + + err = ibmebus_create_devices(ibmebus_matches); + if (err) { + device_unregister(&ibmebus_bus_device); + bus_unregister(&ibmebus_bus_type); + return err; + } + + return 0; +} +machine_postcore_initcall(pseries, ibmebus_bus_init); diff --git a/arch/powerpc/platforms/pseries/io_event_irq.c b/arch/powerpc/platforms/pseries/io_event_irq.c new file mode 100644 index 000000000..f411d4fe7 --- /dev/null +++ b/arch/powerpc/platforms/pseries/io_event_irq.c @@ -0,0 +1,161 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright 2010 2011 Mark Nelson and Tseng-Hui (Frank) Lin, IBM Corporation + */ + +#include <linux/errno.h> +#include <linux/slab.h> +#include <linux/export.h> +#include <linux/irq.h> +#include <linux/interrupt.h> +#include <linux/of.h> +#include <linux/list.h> +#include <linux/notifier.h> + +#include <asm/machdep.h> +#include <asm/rtas.h> +#include <asm/irq.h> +#include <asm/io_event_irq.h> + +#include "pseries.h" + +/* + * IO event interrupt is a mechanism provided by RTAS to return + * information about hardware error and non-error events. Device + * drivers can register their event handlers to receive events. + * Device drivers are expected to use atomic_notifier_chain_register() + * and atomic_notifier_chain_unregister() to register and unregister + * their event handlers. Since multiple IO event types and scopes + * share an IO event interrupt, the event handlers are called one + * by one until the IO event is claimed by one of the handlers. + * The event handlers are expected to return NOTIFY_OK if the + * event is handled by the event handler or NOTIFY_DONE if the + * event does not belong to the handler. + * + * Usage: + * + * Notifier function: + * #include <asm/io_event_irq.h> + * int event_handler(struct notifier_block *nb, unsigned long val, void *data) { + * p = (struct pseries_io_event_sect_data *) data; + * if (! is_my_event(p->scope, p->event_type)) return NOTIFY_DONE; + * : + * : + * return NOTIFY_OK; + * } + * struct notifier_block event_nb = { + * .notifier_call = event_handler, + * } + * + * Registration: + * atomic_notifier_chain_register(&pseries_ioei_notifier_list, &event_nb); + * + * Unregistration: + * atomic_notifier_chain_unregister(&pseries_ioei_notifier_list, &event_nb); + */ + +ATOMIC_NOTIFIER_HEAD(pseries_ioei_notifier_list); +EXPORT_SYMBOL_GPL(pseries_ioei_notifier_list); + +static int ioei_check_exception_token; + +static char ioei_rtas_buf[RTAS_DATA_BUF_SIZE] __cacheline_aligned; + +/** + * Find the data portion of an IO Event section from event log. + * @elog: RTAS error/event log. + * + * Return: + * pointer to a valid IO event section data. NULL if not found. + */ +static struct pseries_io_event * ioei_find_event(struct rtas_error_log *elog) +{ + struct pseries_errorlog *sect; + + /* We should only ever get called for io-event interrupts, but if + * we do get called for another type then something went wrong so + * make some noise about it. + * RTAS_TYPE_IO only exists in extended event log version 6 or later. + * No need to check event log version. + */ + if (unlikely(rtas_error_type(elog) != RTAS_TYPE_IO)) { + printk_once(KERN_WARNING"io_event_irq: Unexpected event type %d", + rtas_error_type(elog)); + return NULL; + } + + sect = get_pseries_errorlog(elog, PSERIES_ELOG_SECT_ID_IO_EVENT); + if (unlikely(!sect)) { + printk_once(KERN_WARNING "io_event_irq: RTAS extended event " + "log does not contain an IO Event section. " + "Could be a bug in system firmware!\n"); + return NULL; + } + return (struct pseries_io_event *) §->data; +} + +/* + * PAPR: + * - check-exception returns the first found error or event and clear that + * error or event so it is reported once. + * - Each interrupt returns one event. If a plateform chooses to report + * multiple events through a single interrupt, it must ensure that the + * interrupt remains asserted until check-exception has been used to + * process all out-standing events for that interrupt. + * + * Implementation notes: + * - Events must be processed in the order they are returned. Hence, + * sequential in nature. + * - The owner of an event is determined by combinations of scope, + * event type, and sub-type. There is no easy way to pre-sort clients + * by scope or event type alone. For example, Torrent ISR route change + * event is reported with scope 0x00 (Not Applicable) rather than + * 0x3B (Torrent-hub). It is better to let the clients to identify + * who owns the event. + */ + +static irqreturn_t ioei_interrupt(int irq, void *dev_id) +{ + struct pseries_io_event *event; + int rtas_rc; + + for (;;) { + rtas_rc = rtas_call(ioei_check_exception_token, 6, 1, NULL, + RTAS_VECTOR_EXTERNAL_INTERRUPT, + virq_to_hw(irq), + RTAS_IO_EVENTS, 1 /* Time Critical */, + __pa(ioei_rtas_buf), + RTAS_DATA_BUF_SIZE); + if (rtas_rc != 0) + break; + + event = ioei_find_event((struct rtas_error_log *)ioei_rtas_buf); + if (!event) + continue; + + atomic_notifier_call_chain(&pseries_ioei_notifier_list, + 0, event); + } + return IRQ_HANDLED; +} + +static int __init ioei_init(void) +{ + struct device_node *np; + + ioei_check_exception_token = rtas_function_token(RTAS_FN_CHECK_EXCEPTION); + if (ioei_check_exception_token == RTAS_UNKNOWN_SERVICE) + return -ENODEV; + + np = of_find_node_by_path("/event-sources/ibm,io-events"); + if (np) { + request_event_sources_irqs(np, ioei_interrupt, "IO_EVENT"); + pr_info("IBM I/O event interrupts enabled\n"); + of_node_put(np); + } else { + return -ENODEV; + } + return 0; +} +machine_subsys_initcall(pseries, ioei_init); + diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c new file mode 100644 index 000000000..496e16c58 --- /dev/null +++ b/arch/powerpc/platforms/pseries/iommu.c @@ -0,0 +1,1742 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2001 Mike Corrigan & Dave Engebretsen, IBM Corporation + * + * Rewrite, cleanup: + * + * Copyright (C) 2004 Olof Johansson <olof@lixom.net>, IBM Corporation + * Copyright (C) 2006 Olof Johansson <olof@lixom.net> + * + * Dynamic DMA mapping support, pSeries-specific parts, both SMP and LPAR. + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/slab.h> +#include <linux/mm.h> +#include <linux/memblock.h> +#include <linux/spinlock.h> +#include <linux/string.h> +#include <linux/pci.h> +#include <linux/dma-mapping.h> +#include <linux/crash_dump.h> +#include <linux/memory.h> +#include <linux/of.h> +#include <linux/of_address.h> +#include <linux/iommu.h> +#include <linux/rculist.h> +#include <asm/io.h> +#include <asm/prom.h> +#include <asm/rtas.h> +#include <asm/iommu.h> +#include <asm/pci-bridge.h> +#include <asm/machdep.h> +#include <asm/firmware.h> +#include <asm/tce.h> +#include <asm/ppc-pci.h> +#include <asm/udbg.h> +#include <asm/mmzone.h> +#include <asm/plpar_wrappers.h> + +#include "pseries.h" + +enum { + DDW_QUERY_PE_DMA_WIN = 0, + DDW_CREATE_PE_DMA_WIN = 1, + DDW_REMOVE_PE_DMA_WIN = 2, + + DDW_APPLICABLE_SIZE +}; + +enum { + DDW_EXT_SIZE = 0, + DDW_EXT_RESET_DMA_WIN = 1, + DDW_EXT_QUERY_OUT_SIZE = 2 +}; + +static struct iommu_table *iommu_pseries_alloc_table(int node) +{ + struct iommu_table *tbl; + + tbl = kzalloc_node(sizeof(struct iommu_table), GFP_KERNEL, node); + if (!tbl) + return NULL; + + INIT_LIST_HEAD_RCU(&tbl->it_group_list); + kref_init(&tbl->it_kref); + return tbl; +} + +static struct iommu_table_group *iommu_pseries_alloc_group(int node) +{ + struct iommu_table_group *table_group; + + table_group = kzalloc_node(sizeof(*table_group), GFP_KERNEL, node); + if (!table_group) + return NULL; + +#ifdef CONFIG_IOMMU_API + table_group->ops = &spapr_tce_table_group_ops; + table_group->pgsizes = SZ_4K; +#endif + + table_group->tables[0] = iommu_pseries_alloc_table(node); + if (table_group->tables[0]) + return table_group; + + kfree(table_group); + return NULL; +} + +static void iommu_pseries_free_group(struct iommu_table_group *table_group, + const char *node_name) +{ + if (!table_group) + return; + +#ifdef CONFIG_IOMMU_API + if (table_group->group) { + iommu_group_put(table_group->group); + BUG_ON(table_group->group); + } +#endif + + /* Default DMA window table is at index 0, while DDW at 1. SR-IOV + * adapters only have table on index 1. + */ + if (table_group->tables[0]) + iommu_tce_table_put(table_group->tables[0]); + + if (table_group->tables[1]) + iommu_tce_table_put(table_group->tables[1]); + + kfree(table_group); +} + +static int tce_build_pSeries(struct iommu_table *tbl, long index, + long npages, unsigned long uaddr, + enum dma_data_direction direction, + unsigned long attrs) +{ + u64 proto_tce; + __be64 *tcep; + u64 rpn; + const unsigned long tceshift = tbl->it_page_shift; + const unsigned long pagesize = IOMMU_PAGE_SIZE(tbl); + + proto_tce = TCE_PCI_READ; // Read allowed + + if (direction != DMA_TO_DEVICE) + proto_tce |= TCE_PCI_WRITE; + + tcep = ((__be64 *)tbl->it_base) + index; + + while (npages--) { + /* can't move this out since we might cross MEMBLOCK boundary */ + rpn = __pa(uaddr) >> tceshift; + *tcep = cpu_to_be64(proto_tce | rpn << tceshift); + + uaddr += pagesize; + tcep++; + } + return 0; +} + + +static void tce_free_pSeries(struct iommu_table *tbl, long index, long npages) +{ + __be64 *tcep; + + tcep = ((__be64 *)tbl->it_base) + index; + + while (npages--) + *(tcep++) = 0; +} + +static unsigned long tce_get_pseries(struct iommu_table *tbl, long index) +{ + __be64 *tcep; + + tcep = ((__be64 *)tbl->it_base) + index; + + return be64_to_cpu(*tcep); +} + +static void tce_free_pSeriesLP(unsigned long liobn, long, long, long); +static void tce_freemulti_pSeriesLP(struct iommu_table*, long, long); + +static int tce_build_pSeriesLP(unsigned long liobn, long tcenum, long tceshift, + long npages, unsigned long uaddr, + enum dma_data_direction direction, + unsigned long attrs) +{ + u64 rc = 0; + u64 proto_tce, tce; + u64 rpn; + int ret = 0; + long tcenum_start = tcenum, npages_start = npages; + + rpn = __pa(uaddr) >> tceshift; + proto_tce = TCE_PCI_READ; + if (direction != DMA_TO_DEVICE) + proto_tce |= TCE_PCI_WRITE; + + while (npages--) { + tce = proto_tce | rpn << tceshift; + rc = plpar_tce_put((u64)liobn, (u64)tcenum << tceshift, tce); + + if (unlikely(rc == H_NOT_ENOUGH_RESOURCES)) { + ret = (int)rc; + tce_free_pSeriesLP(liobn, tcenum_start, tceshift, + (npages_start - (npages + 1))); + break; + } + + if (rc && printk_ratelimit()) { + printk("tce_build_pSeriesLP: plpar_tce_put failed. rc=%lld\n", rc); + printk("\tindex = 0x%llx\n", (u64)liobn); + printk("\ttcenum = 0x%llx\n", (u64)tcenum); + printk("\ttce val = 0x%llx\n", tce ); + dump_stack(); + } + + tcenum++; + rpn++; + } + return ret; +} + +static DEFINE_PER_CPU(__be64 *, tce_page); + +static int tce_buildmulti_pSeriesLP(struct iommu_table *tbl, long tcenum, + long npages, unsigned long uaddr, + enum dma_data_direction direction, + unsigned long attrs) +{ + u64 rc = 0; + u64 proto_tce; + __be64 *tcep; + u64 rpn; + long l, limit; + long tcenum_start = tcenum, npages_start = npages; + int ret = 0; + unsigned long flags; + const unsigned long tceshift = tbl->it_page_shift; + + if ((npages == 1) || !firmware_has_feature(FW_FEATURE_PUT_TCE_IND)) { + return tce_build_pSeriesLP(tbl->it_index, tcenum, + tceshift, npages, uaddr, + direction, attrs); + } + + local_irq_save(flags); /* to protect tcep and the page behind it */ + + tcep = __this_cpu_read(tce_page); + + /* This is safe to do since interrupts are off when we're called + * from iommu_alloc{,_sg}() + */ + if (!tcep) { + tcep = (__be64 *)__get_free_page(GFP_ATOMIC); + /* If allocation fails, fall back to the loop implementation */ + if (!tcep) { + local_irq_restore(flags); + return tce_build_pSeriesLP(tbl->it_index, tcenum, + tceshift, + npages, uaddr, direction, attrs); + } + __this_cpu_write(tce_page, tcep); + } + + rpn = __pa(uaddr) >> tceshift; + proto_tce = TCE_PCI_READ; + if (direction != DMA_TO_DEVICE) + proto_tce |= TCE_PCI_WRITE; + + /* We can map max one pageful of TCEs at a time */ + do { + /* + * Set up the page with TCE data, looping through and setting + * the values. + */ + limit = min_t(long, npages, 4096 / TCE_ENTRY_SIZE); + + for (l = 0; l < limit; l++) { + tcep[l] = cpu_to_be64(proto_tce | rpn << tceshift); + rpn++; + } + + rc = plpar_tce_put_indirect((u64)tbl->it_index, + (u64)tcenum << tceshift, + (u64)__pa(tcep), + limit); + + npages -= limit; + tcenum += limit; + } while (npages > 0 && !rc); + + local_irq_restore(flags); + + if (unlikely(rc == H_NOT_ENOUGH_RESOURCES)) { + ret = (int)rc; + tce_freemulti_pSeriesLP(tbl, tcenum_start, + (npages_start - (npages + limit))); + return ret; + } + + if (rc && printk_ratelimit()) { + printk("tce_buildmulti_pSeriesLP: plpar_tce_put failed. rc=%lld\n", rc); + printk("\tindex = 0x%llx\n", (u64)tbl->it_index); + printk("\tnpages = 0x%llx\n", (u64)npages); + printk("\ttce[0] val = 0x%llx\n", tcep[0]); + dump_stack(); + } + return ret; +} + +static void tce_free_pSeriesLP(unsigned long liobn, long tcenum, long tceshift, + long npages) +{ + u64 rc; + + while (npages--) { + rc = plpar_tce_put((u64)liobn, (u64)tcenum << tceshift, 0); + + if (rc && printk_ratelimit()) { + printk("tce_free_pSeriesLP: plpar_tce_put failed. rc=%lld\n", rc); + printk("\tindex = 0x%llx\n", (u64)liobn); + printk("\ttcenum = 0x%llx\n", (u64)tcenum); + dump_stack(); + } + + tcenum++; + } +} + + +static void tce_freemulti_pSeriesLP(struct iommu_table *tbl, long tcenum, long npages) +{ + u64 rc; + long rpages = npages; + unsigned long limit; + + if (!firmware_has_feature(FW_FEATURE_STUFF_TCE)) + return tce_free_pSeriesLP(tbl->it_index, tcenum, + tbl->it_page_shift, npages); + + do { + limit = min_t(unsigned long, rpages, 512); + + rc = plpar_tce_stuff((u64)tbl->it_index, + (u64)tcenum << tbl->it_page_shift, 0, limit); + + rpages -= limit; + tcenum += limit; + } while (rpages > 0 && !rc); + + if (rc && printk_ratelimit()) { + printk("tce_freemulti_pSeriesLP: plpar_tce_stuff failed\n"); + printk("\trc = %lld\n", rc); + printk("\tindex = 0x%llx\n", (u64)tbl->it_index); + printk("\tnpages = 0x%llx\n", (u64)npages); + dump_stack(); + } +} + +static unsigned long tce_get_pSeriesLP(struct iommu_table *tbl, long tcenum) +{ + u64 rc; + unsigned long tce_ret; + + rc = plpar_tce_get((u64)tbl->it_index, + (u64)tcenum << tbl->it_page_shift, &tce_ret); + + if (rc && printk_ratelimit()) { + printk("tce_get_pSeriesLP: plpar_tce_get failed. rc=%lld\n", rc); + printk("\tindex = 0x%llx\n", (u64)tbl->it_index); + printk("\ttcenum = 0x%llx\n", (u64)tcenum); + dump_stack(); + } + + return tce_ret; +} + +/* this is compatible with cells for the device tree property */ +struct dynamic_dma_window_prop { + __be32 liobn; /* tce table number */ + __be64 dma_base; /* address hi,lo */ + __be32 tce_shift; /* ilog2(tce_page_size) */ + __be32 window_shift; /* ilog2(tce_window_size) */ +}; + +struct dma_win { + struct device_node *device; + const struct dynamic_dma_window_prop *prop; + bool direct; + struct list_head list; +}; + +/* Dynamic DMA Window support */ +struct ddw_query_response { + u32 windows_available; + u64 largest_available_block; + u32 page_size; + u32 migration_capable; +}; + +struct ddw_create_response { + u32 liobn; + u32 addr_hi; + u32 addr_lo; +}; + +static LIST_HEAD(dma_win_list); +/* prevents races between memory on/offline and window creation */ +static DEFINE_SPINLOCK(dma_win_list_lock); +/* protects initializing window twice for same device */ +static DEFINE_MUTEX(dma_win_init_mutex); + +static int tce_clearrange_multi_pSeriesLP(unsigned long start_pfn, + unsigned long num_pfn, const void *arg) +{ + const struct dynamic_dma_window_prop *maprange = arg; + int rc; + u64 tce_size, num_tce, dma_offset, next; + u32 tce_shift; + long limit; + + tce_shift = be32_to_cpu(maprange->tce_shift); + tce_size = 1ULL << tce_shift; + next = start_pfn << PAGE_SHIFT; + num_tce = num_pfn << PAGE_SHIFT; + + /* round back to the beginning of the tce page size */ + num_tce += next & (tce_size - 1); + next &= ~(tce_size - 1); + + /* covert to number of tces */ + num_tce |= tce_size - 1; + num_tce >>= tce_shift; + + do { + /* + * Set up the page with TCE data, looping through and setting + * the values. + */ + limit = min_t(long, num_tce, 512); + dma_offset = next + be64_to_cpu(maprange->dma_base); + + rc = plpar_tce_stuff((u64)be32_to_cpu(maprange->liobn), + dma_offset, + 0, limit); + next += limit * tce_size; + num_tce -= limit; + } while (num_tce > 0 && !rc); + + return rc; +} + +static int tce_setrange_multi_pSeriesLP(unsigned long start_pfn, + unsigned long num_pfn, const void *arg) +{ + const struct dynamic_dma_window_prop *maprange = arg; + u64 tce_size, num_tce, dma_offset, next, proto_tce, liobn; + __be64 *tcep; + u32 tce_shift; + u64 rc = 0; + long l, limit; + + if (!firmware_has_feature(FW_FEATURE_PUT_TCE_IND)) { + unsigned long tceshift = be32_to_cpu(maprange->tce_shift); + unsigned long dmastart = (start_pfn << PAGE_SHIFT) + + be64_to_cpu(maprange->dma_base); + unsigned long tcenum = dmastart >> tceshift; + unsigned long npages = num_pfn << PAGE_SHIFT >> tceshift; + void *uaddr = __va(start_pfn << PAGE_SHIFT); + + return tce_build_pSeriesLP(be32_to_cpu(maprange->liobn), + tcenum, tceshift, npages, (unsigned long) uaddr, + DMA_BIDIRECTIONAL, 0); + } + + local_irq_disable(); /* to protect tcep and the page behind it */ + tcep = __this_cpu_read(tce_page); + + if (!tcep) { + tcep = (__be64 *)__get_free_page(GFP_ATOMIC); + if (!tcep) { + local_irq_enable(); + return -ENOMEM; + } + __this_cpu_write(tce_page, tcep); + } + + proto_tce = TCE_PCI_READ | TCE_PCI_WRITE; + + liobn = (u64)be32_to_cpu(maprange->liobn); + tce_shift = be32_to_cpu(maprange->tce_shift); + tce_size = 1ULL << tce_shift; + next = start_pfn << PAGE_SHIFT; + num_tce = num_pfn << PAGE_SHIFT; + + /* round back to the beginning of the tce page size */ + num_tce += next & (tce_size - 1); + next &= ~(tce_size - 1); + + /* covert to number of tces */ + num_tce |= tce_size - 1; + num_tce >>= tce_shift; + + /* We can map max one pageful of TCEs at a time */ + do { + /* + * Set up the page with TCE data, looping through and setting + * the values. + */ + limit = min_t(long, num_tce, 4096 / TCE_ENTRY_SIZE); + dma_offset = next + be64_to_cpu(maprange->dma_base); + + for (l = 0; l < limit; l++) { + tcep[l] = cpu_to_be64(proto_tce | next); + next += tce_size; + } + + rc = plpar_tce_put_indirect(liobn, + dma_offset, + (u64)__pa(tcep), + limit); + + num_tce -= limit; + } while (num_tce > 0 && !rc); + + /* error cleanup: caller will clear whole range */ + + local_irq_enable(); + return rc; +} + +static int tce_setrange_multi_pSeriesLP_walk(unsigned long start_pfn, + unsigned long num_pfn, void *arg) +{ + return tce_setrange_multi_pSeriesLP(start_pfn, num_pfn, arg); +} + +static void iommu_table_setparms_common(struct iommu_table *tbl, unsigned long busno, + unsigned long liobn, unsigned long win_addr, + unsigned long window_size, unsigned long page_shift, + void *base, struct iommu_table_ops *table_ops) +{ + tbl->it_busno = busno; + tbl->it_index = liobn; + tbl->it_offset = win_addr >> page_shift; + tbl->it_size = window_size >> page_shift; + tbl->it_page_shift = page_shift; + tbl->it_base = (unsigned long)base; + tbl->it_blocksize = 16; + tbl->it_type = TCE_PCI; + tbl->it_ops = table_ops; +} + +struct iommu_table_ops iommu_table_pseries_ops; + +static void iommu_table_setparms(struct pci_controller *phb, + struct device_node *dn, + struct iommu_table *tbl) +{ + struct device_node *node; + const unsigned long *basep; + const u32 *sizep; + + /* Test if we are going over 2GB of DMA space */ + if (phb->dma_window_base_cur + phb->dma_window_size > SZ_2G) { + udbg_printf("PCI_DMA: Unexpected number of IOAs under this PHB.\n"); + panic("PCI_DMA: Unexpected number of IOAs under this PHB.\n"); + } + + node = phb->dn; + basep = of_get_property(node, "linux,tce-base", NULL); + sizep = of_get_property(node, "linux,tce-size", NULL); + if (basep == NULL || sizep == NULL) { + printk(KERN_ERR "PCI_DMA: iommu_table_setparms: %pOF has " + "missing tce entries !\n", dn); + return; + } + + iommu_table_setparms_common(tbl, phb->bus->number, 0, phb->dma_window_base_cur, + phb->dma_window_size, IOMMU_PAGE_SHIFT_4K, + __va(*basep), &iommu_table_pseries_ops); + + if (!is_kdump_kernel()) + memset((void *)tbl->it_base, 0, *sizep); + + phb->dma_window_base_cur += phb->dma_window_size; +} + +struct iommu_table_ops iommu_table_lpar_multi_ops; + +/* + * iommu_table_setparms_lpar + * + * Function: On pSeries LPAR systems, return TCE table info, given a pci bus. + */ +static void iommu_table_setparms_lpar(struct pci_controller *phb, + struct device_node *dn, + struct iommu_table *tbl, + struct iommu_table_group *table_group, + const __be32 *dma_window) +{ + unsigned long offset, size, liobn; + + of_parse_dma_window(dn, dma_window, &liobn, &offset, &size); + + iommu_table_setparms_common(tbl, phb->bus->number, liobn, offset, size, IOMMU_PAGE_SHIFT_4K, NULL, + &iommu_table_lpar_multi_ops); + + + table_group->tce32_start = offset; + table_group->tce32_size = size; +} + +struct iommu_table_ops iommu_table_pseries_ops = { + .set = tce_build_pSeries, + .clear = tce_free_pSeries, + .get = tce_get_pseries +}; + +static void pci_dma_bus_setup_pSeries(struct pci_bus *bus) +{ + struct device_node *dn; + struct iommu_table *tbl; + struct device_node *isa_dn, *isa_dn_orig; + struct device_node *tmp; + struct pci_dn *pci; + int children; + + dn = pci_bus_to_OF_node(bus); + + pr_debug("pci_dma_bus_setup_pSeries: setting up bus %pOF\n", dn); + + if (bus->self) { + /* This is not a root bus, any setup will be done for the + * device-side of the bridge in iommu_dev_setup_pSeries(). + */ + return; + } + pci = PCI_DN(dn); + + /* Check if the ISA bus on the system is under + * this PHB. + */ + isa_dn = isa_dn_orig = of_find_node_by_type(NULL, "isa"); + + while (isa_dn && isa_dn != dn) + isa_dn = isa_dn->parent; + + of_node_put(isa_dn_orig); + + /* Count number of direct PCI children of the PHB. */ + for (children = 0, tmp = dn->child; tmp; tmp = tmp->sibling) + children++; + + pr_debug("Children: %d\n", children); + + /* Calculate amount of DMA window per slot. Each window must be + * a power of two (due to pci_alloc_consistent requirements). + * + * Keep 256MB aside for PHBs with ISA. + */ + + if (!isa_dn) { + /* No ISA/IDE - just set window size and return */ + pci->phb->dma_window_size = 0x80000000ul; /* To be divided */ + + while (pci->phb->dma_window_size * children > 0x80000000ul) + pci->phb->dma_window_size >>= 1; + pr_debug("No ISA/IDE, window size is 0x%llx\n", + pci->phb->dma_window_size); + pci->phb->dma_window_base_cur = 0; + + return; + } + + /* If we have ISA, then we probably have an IDE + * controller too. Allocate a 128MB table but + * skip the first 128MB to avoid stepping on ISA + * space. + */ + pci->phb->dma_window_size = 0x8000000ul; + pci->phb->dma_window_base_cur = 0x8000000ul; + + pci->table_group = iommu_pseries_alloc_group(pci->phb->node); + tbl = pci->table_group->tables[0]; + + iommu_table_setparms(pci->phb, dn, tbl); + + if (!iommu_init_table(tbl, pci->phb->node, 0, 0)) + panic("Failed to initialize iommu table"); + + /* Divide the rest (1.75GB) among the children */ + pci->phb->dma_window_size = 0x80000000ul; + while (pci->phb->dma_window_size * children > 0x70000000ul) + pci->phb->dma_window_size >>= 1; + + pr_debug("ISA/IDE, window size is 0x%llx\n", pci->phb->dma_window_size); +} + +#ifdef CONFIG_IOMMU_API +static int tce_exchange_pseries(struct iommu_table *tbl, long index, unsigned + long *tce, enum dma_data_direction *direction) +{ + long rc; + unsigned long ioba = (unsigned long) index << tbl->it_page_shift; + unsigned long flags, oldtce = 0; + u64 proto_tce = iommu_direction_to_tce_perm(*direction); + unsigned long newtce = *tce | proto_tce; + + spin_lock_irqsave(&tbl->large_pool.lock, flags); + + rc = plpar_tce_get((u64)tbl->it_index, ioba, &oldtce); + if (!rc) + rc = plpar_tce_put((u64)tbl->it_index, ioba, newtce); + + if (!rc) { + *direction = iommu_tce_direction(oldtce); + *tce = oldtce & ~(TCE_PCI_READ | TCE_PCI_WRITE); + } + + spin_unlock_irqrestore(&tbl->large_pool.lock, flags); + + return rc; +} +#endif + +struct iommu_table_ops iommu_table_lpar_multi_ops = { + .set = tce_buildmulti_pSeriesLP, +#ifdef CONFIG_IOMMU_API + .xchg_no_kill = tce_exchange_pseries, +#endif + .clear = tce_freemulti_pSeriesLP, + .get = tce_get_pSeriesLP +}; + +/* + * Find nearest ibm,dma-window (default DMA window) or direct DMA window or + * dynamic 64bit DMA window, walking up the device tree. + */ +static struct device_node *pci_dma_find(struct device_node *dn, + const __be32 **dma_window) +{ + const __be32 *dw = NULL; + + for ( ; dn && PCI_DN(dn); dn = dn->parent) { + dw = of_get_property(dn, "ibm,dma-window", NULL); + if (dw) { + if (dma_window) + *dma_window = dw; + return dn; + } + dw = of_get_property(dn, DIRECT64_PROPNAME, NULL); + if (dw) + return dn; + dw = of_get_property(dn, DMA64_PROPNAME, NULL); + if (dw) + return dn; + } + + return NULL; +} + +static void pci_dma_bus_setup_pSeriesLP(struct pci_bus *bus) +{ + struct iommu_table *tbl; + struct device_node *dn, *pdn; + struct pci_dn *ppci; + const __be32 *dma_window = NULL; + + dn = pci_bus_to_OF_node(bus); + + pr_debug("pci_dma_bus_setup_pSeriesLP: setting up bus %pOF\n", + dn); + + pdn = pci_dma_find(dn, &dma_window); + + if (dma_window == NULL) + pr_debug(" no ibm,dma-window property !\n"); + + ppci = PCI_DN(pdn); + + pr_debug(" parent is %pOF, iommu_table: 0x%p\n", + pdn, ppci->table_group); + + if (!ppci->table_group) { + ppci->table_group = iommu_pseries_alloc_group(ppci->phb->node); + tbl = ppci->table_group->tables[0]; + if (dma_window) { + iommu_table_setparms_lpar(ppci->phb, pdn, tbl, + ppci->table_group, dma_window); + + if (!iommu_init_table(tbl, ppci->phb->node, 0, 0)) + panic("Failed to initialize iommu table"); + } + iommu_register_group(ppci->table_group, + pci_domain_nr(bus), 0); + pr_debug(" created table: %p\n", ppci->table_group); + } +} + + +static void pci_dma_dev_setup_pSeries(struct pci_dev *dev) +{ + struct device_node *dn; + struct iommu_table *tbl; + + pr_debug("pci_dma_dev_setup_pSeries: %s\n", pci_name(dev)); + + dn = dev->dev.of_node; + + /* If we're the direct child of a root bus, then we need to allocate + * an iommu table ourselves. The bus setup code should have setup + * the window sizes already. + */ + if (!dev->bus->self) { + struct pci_controller *phb = PCI_DN(dn)->phb; + + pr_debug(" --> first child, no bridge. Allocating iommu table.\n"); + PCI_DN(dn)->table_group = iommu_pseries_alloc_group(phb->node); + tbl = PCI_DN(dn)->table_group->tables[0]; + iommu_table_setparms(phb, dn, tbl); + + if (!iommu_init_table(tbl, phb->node, 0, 0)) + panic("Failed to initialize iommu table"); + + set_iommu_table_base(&dev->dev, tbl); + return; + } + + /* If this device is further down the bus tree, search upwards until + * an already allocated iommu table is found and use that. + */ + + while (dn && PCI_DN(dn) && PCI_DN(dn)->table_group == NULL) + dn = dn->parent; + + if (dn && PCI_DN(dn)) + set_iommu_table_base(&dev->dev, + PCI_DN(dn)->table_group->tables[0]); + else + printk(KERN_WARNING "iommu: Device %s has no iommu table\n", + pci_name(dev)); +} + +static int __read_mostly disable_ddw; + +static int __init disable_ddw_setup(char *str) +{ + disable_ddw = 1; + printk(KERN_INFO "ppc iommu: disabling ddw.\n"); + + return 0; +} + +early_param("disable_ddw", disable_ddw_setup); + +static void clean_dma_window(struct device_node *np, struct dynamic_dma_window_prop *dwp) +{ + int ret; + + ret = tce_clearrange_multi_pSeriesLP(0, + 1ULL << (be32_to_cpu(dwp->window_shift) - PAGE_SHIFT), dwp); + if (ret) + pr_warn("%pOF failed to clear tces in window.\n", + np); + else + pr_debug("%pOF successfully cleared tces in window.\n", + np); +} + +/* + * Call only if DMA window is clean. + */ +static void __remove_dma_window(struct device_node *np, u32 *ddw_avail, u64 liobn) +{ + int ret; + + ret = rtas_call(ddw_avail[DDW_REMOVE_PE_DMA_WIN], 1, 1, NULL, liobn); + if (ret) + pr_warn("%pOF: failed to remove DMA window: rtas returned " + "%d to ibm,remove-pe-dma-window(%x) %llx\n", + np, ret, ddw_avail[DDW_REMOVE_PE_DMA_WIN], liobn); + else + pr_debug("%pOF: successfully removed DMA window: rtas returned " + "%d to ibm,remove-pe-dma-window(%x) %llx\n", + np, ret, ddw_avail[DDW_REMOVE_PE_DMA_WIN], liobn); +} + +static void remove_dma_window(struct device_node *np, u32 *ddw_avail, + struct property *win) +{ + struct dynamic_dma_window_prop *dwp; + u64 liobn; + + dwp = win->value; + liobn = (u64)be32_to_cpu(dwp->liobn); + + clean_dma_window(np, dwp); + __remove_dma_window(np, ddw_avail, liobn); +} + +static int remove_ddw(struct device_node *np, bool remove_prop, const char *win_name) +{ + struct property *win; + u32 ddw_avail[DDW_APPLICABLE_SIZE]; + int ret = 0; + + win = of_find_property(np, win_name, NULL); + if (!win) + return -EINVAL; + + ret = of_property_read_u32_array(np, "ibm,ddw-applicable", + &ddw_avail[0], DDW_APPLICABLE_SIZE); + if (ret) + return 0; + + + if (win->length >= sizeof(struct dynamic_dma_window_prop)) + remove_dma_window(np, ddw_avail, win); + + if (!remove_prop) + return 0; + + ret = of_remove_property(np, win); + if (ret) + pr_warn("%pOF: failed to remove DMA window property: %d\n", + np, ret); + return 0; +} + +static bool find_existing_ddw(struct device_node *pdn, u64 *dma_addr, int *window_shift, + bool *direct_mapping) +{ + struct dma_win *window; + const struct dynamic_dma_window_prop *dma64; + bool found = false; + + spin_lock(&dma_win_list_lock); + /* check if we already created a window and dupe that config if so */ + list_for_each_entry(window, &dma_win_list, list) { + if (window->device == pdn) { + dma64 = window->prop; + *dma_addr = be64_to_cpu(dma64->dma_base); + *window_shift = be32_to_cpu(dma64->window_shift); + *direct_mapping = window->direct; + found = true; + break; + } + } + spin_unlock(&dma_win_list_lock); + + return found; +} + +static struct dma_win *ddw_list_new_entry(struct device_node *pdn, + const struct dynamic_dma_window_prop *dma64) +{ + struct dma_win *window; + + window = kzalloc(sizeof(*window), GFP_KERNEL); + if (!window) + return NULL; + + window->device = pdn; + window->prop = dma64; + window->direct = false; + + return window; +} + +static void find_existing_ddw_windows_named(const char *name) +{ + int len; + struct device_node *pdn; + struct dma_win *window; + const struct dynamic_dma_window_prop *dma64; + + for_each_node_with_property(pdn, name) { + dma64 = of_get_property(pdn, name, &len); + if (!dma64 || len < sizeof(*dma64)) { + remove_ddw(pdn, true, name); + continue; + } + + window = ddw_list_new_entry(pdn, dma64); + if (!window) { + of_node_put(pdn); + break; + } + + spin_lock(&dma_win_list_lock); + list_add(&window->list, &dma_win_list); + spin_unlock(&dma_win_list_lock); + } +} + +static int find_existing_ddw_windows(void) +{ + if (!firmware_has_feature(FW_FEATURE_LPAR)) + return 0; + + find_existing_ddw_windows_named(DIRECT64_PROPNAME); + find_existing_ddw_windows_named(DMA64_PROPNAME); + + return 0; +} +machine_arch_initcall(pseries, find_existing_ddw_windows); + +/** + * ddw_read_ext - Get the value of an DDW extension + * @np: device node from which the extension value is to be read. + * @extnum: index number of the extension. + * @value: pointer to return value, modified when extension is available. + * + * Checks if "ibm,ddw-extensions" exists for this node, and get the value + * on index 'extnum'. + * It can be used only to check if a property exists, passing value == NULL. + * + * Returns: + * 0 if extension successfully read + * -EINVAL if the "ibm,ddw-extensions" does not exist, + * -ENODATA if "ibm,ddw-extensions" does not have a value, and + * -EOVERFLOW if "ibm,ddw-extensions" does not contain this extension. + */ +static inline int ddw_read_ext(const struct device_node *np, int extnum, + u32 *value) +{ + static const char propname[] = "ibm,ddw-extensions"; + u32 count; + int ret; + + ret = of_property_read_u32_index(np, propname, DDW_EXT_SIZE, &count); + if (ret) + return ret; + + if (count < extnum) + return -EOVERFLOW; + + if (!value) + value = &count; + + return of_property_read_u32_index(np, propname, extnum, value); +} + +static int query_ddw(struct pci_dev *dev, const u32 *ddw_avail, + struct ddw_query_response *query, + struct device_node *parent) +{ + struct device_node *dn; + struct pci_dn *pdn; + u32 cfg_addr, ext_query, query_out[5]; + u64 buid; + int ret, out_sz; + + /* + * From LoPAR level 2.8, "ibm,ddw-extensions" index 3 can rule how many + * output parameters ibm,query-pe-dma-windows will have, ranging from + * 5 to 6. + */ + ret = ddw_read_ext(parent, DDW_EXT_QUERY_OUT_SIZE, &ext_query); + if (!ret && ext_query == 1) + out_sz = 6; + else + out_sz = 5; + + /* + * Get the config address and phb buid of the PE window. + * Rely on eeh to retrieve this for us. + * Retrieve them from the pci device, not the node with the + * dma-window property + */ + dn = pci_device_to_OF_node(dev); + pdn = PCI_DN(dn); + buid = pdn->phb->buid; + cfg_addr = ((pdn->busno << 16) | (pdn->devfn << 8)); + + ret = rtas_call(ddw_avail[DDW_QUERY_PE_DMA_WIN], 3, out_sz, query_out, + cfg_addr, BUID_HI(buid), BUID_LO(buid)); + + switch (out_sz) { + case 5: + query->windows_available = query_out[0]; + query->largest_available_block = query_out[1]; + query->page_size = query_out[2]; + query->migration_capable = query_out[3]; + break; + case 6: + query->windows_available = query_out[0]; + query->largest_available_block = ((u64)query_out[1] << 32) | + query_out[2]; + query->page_size = query_out[3]; + query->migration_capable = query_out[4]; + break; + } + + dev_info(&dev->dev, "ibm,query-pe-dma-windows(%x) %x %x %x returned %d, lb=%llx ps=%x wn=%d\n", + ddw_avail[DDW_QUERY_PE_DMA_WIN], cfg_addr, BUID_HI(buid), + BUID_LO(buid), ret, query->largest_available_block, + query->page_size, query->windows_available); + + return ret; +} + +static int create_ddw(struct pci_dev *dev, const u32 *ddw_avail, + struct ddw_create_response *create, int page_shift, + int window_shift) +{ + struct device_node *dn; + struct pci_dn *pdn; + u32 cfg_addr; + u64 buid; + int ret; + + /* + * Get the config address and phb buid of the PE window. + * Rely on eeh to retrieve this for us. + * Retrieve them from the pci device, not the node with the + * dma-window property + */ + dn = pci_device_to_OF_node(dev); + pdn = PCI_DN(dn); + buid = pdn->phb->buid; + cfg_addr = ((pdn->busno << 16) | (pdn->devfn << 8)); + + do { + /* extra outputs are LIOBN and dma-addr (hi, lo) */ + ret = rtas_call(ddw_avail[DDW_CREATE_PE_DMA_WIN], 5, 4, + (u32 *)create, cfg_addr, BUID_HI(buid), + BUID_LO(buid), page_shift, window_shift); + } while (rtas_busy_delay(ret)); + dev_info(&dev->dev, + "ibm,create-pe-dma-window(%x) %x %x %x %x %x returned %d " + "(liobn = 0x%x starting addr = %x %x)\n", + ddw_avail[DDW_CREATE_PE_DMA_WIN], cfg_addr, BUID_HI(buid), + BUID_LO(buid), page_shift, window_shift, ret, create->liobn, + create->addr_hi, create->addr_lo); + + return ret; +} + +struct failed_ddw_pdn { + struct device_node *pdn; + struct list_head list; +}; + +static LIST_HEAD(failed_ddw_pdn_list); + +static phys_addr_t ddw_memory_hotplug_max(void) +{ + resource_size_t max_addr = memory_hotplug_max(); + struct device_node *memory; + + for_each_node_by_type(memory, "memory") { + struct resource res; + + if (of_address_to_resource(memory, 0, &res)) + continue; + + max_addr = max_t(resource_size_t, max_addr, res.end + 1); + } + + return max_addr; +} + +/* + * Platforms supporting the DDW option starting with LoPAR level 2.7 implement + * ibm,ddw-extensions, which carries the rtas token for + * ibm,reset-pe-dma-windows. + * That rtas-call can be used to restore the default DMA window for the device. + */ +static void reset_dma_window(struct pci_dev *dev, struct device_node *par_dn) +{ + int ret; + u32 cfg_addr, reset_dma_win; + u64 buid; + struct device_node *dn; + struct pci_dn *pdn; + + ret = ddw_read_ext(par_dn, DDW_EXT_RESET_DMA_WIN, &reset_dma_win); + if (ret) + return; + + dn = pci_device_to_OF_node(dev); + pdn = PCI_DN(dn); + buid = pdn->phb->buid; + cfg_addr = (pdn->busno << 16) | (pdn->devfn << 8); + + ret = rtas_call(reset_dma_win, 3, 1, NULL, cfg_addr, BUID_HI(buid), + BUID_LO(buid)); + if (ret) + dev_info(&dev->dev, + "ibm,reset-pe-dma-windows(%x) %x %x %x returned %d ", + reset_dma_win, cfg_addr, BUID_HI(buid), BUID_LO(buid), + ret); +} + +/* Return largest page shift based on "IO Page Sizes" output of ibm,query-pe-dma-window. */ +static int iommu_get_page_shift(u32 query_page_size) +{ + /* Supported IO page-sizes according to LoPAR, note that 2M is out of order */ + const int shift[] = { + __builtin_ctzll(SZ_4K), __builtin_ctzll(SZ_64K), __builtin_ctzll(SZ_16M), + __builtin_ctzll(SZ_32M), __builtin_ctzll(SZ_64M), __builtin_ctzll(SZ_128M), + __builtin_ctzll(SZ_256M), __builtin_ctzll(SZ_16G), __builtin_ctzll(SZ_2M) + }; + + int i = ARRAY_SIZE(shift) - 1; + int ret = 0; + + /* + * On LoPAR, ibm,query-pe-dma-window outputs "IO Page Sizes" using a bit field: + * - bit 31 means 4k pages are supported, + * - bit 30 means 64k pages are supported, and so on. + * Larger pagesizes map more memory with the same amount of TCEs, so start probing them. + */ + for (; i >= 0 ; i--) { + if (query_page_size & (1 << i)) + ret = max(ret, shift[i]); + } + + return ret; +} + +static struct property *ddw_property_create(const char *propname, u32 liobn, u64 dma_addr, + u32 page_shift, u32 window_shift) +{ + struct dynamic_dma_window_prop *ddwprop; + struct property *win64; + + win64 = kzalloc(sizeof(*win64), GFP_KERNEL); + if (!win64) + return NULL; + + win64->name = kstrdup(propname, GFP_KERNEL); + ddwprop = kzalloc(sizeof(*ddwprop), GFP_KERNEL); + win64->value = ddwprop; + win64->length = sizeof(*ddwprop); + if (!win64->name || !win64->value) { + kfree(win64->name); + kfree(win64->value); + kfree(win64); + return NULL; + } + + ddwprop->liobn = cpu_to_be32(liobn); + ddwprop->dma_base = cpu_to_be64(dma_addr); + ddwprop->tce_shift = cpu_to_be32(page_shift); + ddwprop->window_shift = cpu_to_be32(window_shift); + + return win64; +} + +/* + * If the PE supports dynamic dma windows, and there is space for a table + * that can map all pages in a linear offset, then setup such a table, + * and record the dma-offset in the struct device. + * + * dev: the pci device we are checking + * pdn: the parent pe node with the ibm,dma_window property + * Future: also check if we can remap the base window for our base page size + * + * returns true if can map all pages (direct mapping), false otherwise.. + */ +static bool enable_ddw(struct pci_dev *dev, struct device_node *pdn) +{ + int len = 0, ret; + int max_ram_len = order_base_2(ddw_memory_hotplug_max()); + struct ddw_query_response query; + struct ddw_create_response create; + int page_shift; + u64 win_addr; + const char *win_name; + struct device_node *dn; + u32 ddw_avail[DDW_APPLICABLE_SIZE]; + struct dma_win *window; + struct property *win64; + struct failed_ddw_pdn *fpdn; + bool default_win_removed = false, direct_mapping = false; + bool pmem_present; + struct pci_dn *pci = PCI_DN(pdn); + struct property *default_win = NULL; + + dn = of_find_node_by_type(NULL, "ibm,pmemory"); + pmem_present = dn != NULL; + of_node_put(dn); + + mutex_lock(&dma_win_init_mutex); + + if (find_existing_ddw(pdn, &dev->dev.archdata.dma_offset, &len, &direct_mapping)) + goto out_unlock; + + /* + * If we already went through this for a previous function of + * the same device and failed, we don't want to muck with the + * DMA window again, as it will race with in-flight operations + * and can lead to EEHs. The above mutex protects access to the + * list. + */ + list_for_each_entry(fpdn, &failed_ddw_pdn_list, list) { + if (fpdn->pdn == pdn) + goto out_unlock; + } + + /* + * the ibm,ddw-applicable property holds the tokens for: + * ibm,query-pe-dma-window + * ibm,create-pe-dma-window + * ibm,remove-pe-dma-window + * for the given node in that order. + * the property is actually in the parent, not the PE + */ + ret = of_property_read_u32_array(pdn, "ibm,ddw-applicable", + &ddw_avail[0], DDW_APPLICABLE_SIZE); + if (ret) + goto out_failed; + + /* + * Query if there is a second window of size to map the + * whole partition. Query returns number of windows, largest + * block assigned to PE (partition endpoint), and two bitmasks + * of page sizes: supported and supported for migrate-dma. + */ + dn = pci_device_to_OF_node(dev); + ret = query_ddw(dev, ddw_avail, &query, pdn); + if (ret != 0) + goto out_failed; + + /* + * If there is no window available, remove the default DMA window, + * if it's present. This will make all the resources available to the + * new DDW window. + * If anything fails after this, we need to restore it, so also check + * for extensions presence. + */ + if (query.windows_available == 0) { + int reset_win_ext; + + /* DDW + IOMMU on single window may fail if there is any allocation */ + if (iommu_table_in_use(pci->table_group->tables[0])) { + dev_warn(&dev->dev, "current IOMMU table in use, can't be replaced.\n"); + goto out_failed; + } + + default_win = of_find_property(pdn, "ibm,dma-window", NULL); + if (!default_win) + goto out_failed; + + reset_win_ext = ddw_read_ext(pdn, DDW_EXT_RESET_DMA_WIN, NULL); + if (reset_win_ext) + goto out_failed; + + remove_dma_window(pdn, ddw_avail, default_win); + default_win_removed = true; + + /* Query again, to check if the window is available */ + ret = query_ddw(dev, ddw_avail, &query, pdn); + if (ret != 0) + goto out_failed; + + if (query.windows_available == 0) { + /* no windows are available for this device. */ + dev_dbg(&dev->dev, "no free dynamic windows"); + goto out_failed; + } + } + + page_shift = iommu_get_page_shift(query.page_size); + if (!page_shift) { + dev_dbg(&dev->dev, "no supported page size in mask %x", + query.page_size); + goto out_failed; + } + + + /* + * The "ibm,pmemory" can appear anywhere in the address space. + * Assuming it is still backed by page structs, try MAX_PHYSMEM_BITS + * for the upper limit and fallback to max RAM otherwise but this + * disables device::dma_ops_bypass. + */ + len = max_ram_len; + if (pmem_present) { + if (query.largest_available_block >= + (1ULL << (MAX_PHYSMEM_BITS - page_shift))) + len = MAX_PHYSMEM_BITS; + else + dev_info(&dev->dev, "Skipping ibm,pmemory"); + } + + /* check if the available block * number of ptes will map everything */ + if (query.largest_available_block < (1ULL << (len - page_shift))) { + dev_dbg(&dev->dev, + "can't map partition max 0x%llx with %llu %llu-sized pages\n", + 1ULL << len, + query.largest_available_block, + 1ULL << page_shift); + + len = order_base_2(query.largest_available_block << page_shift); + win_name = DMA64_PROPNAME; + } else { + direct_mapping = !default_win_removed || + (len == MAX_PHYSMEM_BITS) || + (!pmem_present && (len == max_ram_len)); + win_name = direct_mapping ? DIRECT64_PROPNAME : DMA64_PROPNAME; + } + + ret = create_ddw(dev, ddw_avail, &create, page_shift, len); + if (ret != 0) + goto out_failed; + + dev_dbg(&dev->dev, "created tce table LIOBN 0x%x for %pOF\n", + create.liobn, dn); + + win_addr = ((u64)create.addr_hi << 32) | create.addr_lo; + win64 = ddw_property_create(win_name, create.liobn, win_addr, page_shift, len); + + if (!win64) { + dev_info(&dev->dev, + "couldn't allocate property, property name, or value\n"); + goto out_remove_win; + } + + ret = of_add_property(pdn, win64); + if (ret) { + dev_err(&dev->dev, "unable to add DMA window property for %pOF: %d", + pdn, ret); + goto out_free_prop; + } + + window = ddw_list_new_entry(pdn, win64->value); + if (!window) + goto out_del_prop; + + if (direct_mapping) { + window->direct = true; + + /* DDW maps the whole partition, so enable direct DMA mapping */ + ret = walk_system_ram_range(0, memblock_end_of_DRAM() >> PAGE_SHIFT, + win64->value, tce_setrange_multi_pSeriesLP_walk); + if (ret) { + dev_info(&dev->dev, "failed to map DMA window for %pOF: %d\n", + dn, ret); + + /* Make sure to clean DDW if any TCE was set*/ + clean_dma_window(pdn, win64->value); + goto out_del_list; + } + } else { + struct iommu_table *newtbl; + int i; + unsigned long start = 0, end = 0; + + window->direct = false; + + for (i = 0; i < ARRAY_SIZE(pci->phb->mem_resources); i++) { + const unsigned long mask = IORESOURCE_MEM_64 | IORESOURCE_MEM; + + /* Look for MMIO32 */ + if ((pci->phb->mem_resources[i].flags & mask) == IORESOURCE_MEM) { + start = pci->phb->mem_resources[i].start; + end = pci->phb->mem_resources[i].end; + break; + } + } + + /* New table for using DDW instead of the default DMA window */ + newtbl = iommu_pseries_alloc_table(pci->phb->node); + if (!newtbl) { + dev_dbg(&dev->dev, "couldn't create new IOMMU table\n"); + goto out_del_list; + } + + iommu_table_setparms_common(newtbl, pci->phb->bus->number, create.liobn, win_addr, + 1UL << len, page_shift, NULL, &iommu_table_lpar_multi_ops); + iommu_init_table(newtbl, pci->phb->node, start, end); + + pci->table_group->tables[1] = newtbl; + + set_iommu_table_base(&dev->dev, newtbl); + } + + if (default_win_removed) { + iommu_tce_table_put(pci->table_group->tables[0]); + pci->table_group->tables[0] = NULL; + + /* default_win is valid here because default_win_removed == true */ + of_remove_property(pdn, default_win); + dev_info(&dev->dev, "Removed default DMA window for %pOF\n", pdn); + } + + spin_lock(&dma_win_list_lock); + list_add(&window->list, &dma_win_list); + spin_unlock(&dma_win_list_lock); + + dev->dev.archdata.dma_offset = win_addr; + goto out_unlock; + +out_del_list: + kfree(window); + +out_del_prop: + of_remove_property(pdn, win64); + +out_free_prop: + kfree(win64->name); + kfree(win64->value); + kfree(win64); + +out_remove_win: + /* DDW is clean, so it's ok to call this directly. */ + __remove_dma_window(pdn, ddw_avail, create.liobn); + +out_failed: + if (default_win_removed) + reset_dma_window(dev, pdn); + + fpdn = kzalloc(sizeof(*fpdn), GFP_KERNEL); + if (!fpdn) + goto out_unlock; + fpdn->pdn = pdn; + list_add(&fpdn->list, &failed_ddw_pdn_list); + +out_unlock: + mutex_unlock(&dma_win_init_mutex); + + /* + * If we have persistent memory and the window size is only as big + * as RAM, then we failed to create a window to cover persistent + * memory and need to set the DMA limit. + */ + if (pmem_present && direct_mapping && len == max_ram_len) + dev->dev.bus_dma_limit = dev->dev.archdata.dma_offset + (1ULL << len); + + return direct_mapping; +} + +static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev) +{ + struct device_node *pdn, *dn; + struct iommu_table *tbl; + const __be32 *dma_window = NULL; + struct pci_dn *pci; + + pr_debug("pci_dma_dev_setup_pSeriesLP: %s\n", pci_name(dev)); + + /* dev setup for LPAR is a little tricky, since the device tree might + * contain the dma-window properties per-device and not necessarily + * for the bus. So we need to search upwards in the tree until we + * either hit a dma-window property, OR find a parent with a table + * already allocated. + */ + dn = pci_device_to_OF_node(dev); + pr_debug(" node is %pOF\n", dn); + + pdn = pci_dma_find(dn, &dma_window); + if (!pdn || !PCI_DN(pdn)) { + printk(KERN_WARNING "pci_dma_dev_setup_pSeriesLP: " + "no DMA window found for pci dev=%s dn=%pOF\n", + pci_name(dev), dn); + return; + } + pr_debug(" parent is %pOF\n", pdn); + + pci = PCI_DN(pdn); + if (!pci->table_group) { + pci->table_group = iommu_pseries_alloc_group(pci->phb->node); + tbl = pci->table_group->tables[0]; + iommu_table_setparms_lpar(pci->phb, pdn, tbl, + pci->table_group, dma_window); + + iommu_init_table(tbl, pci->phb->node, 0, 0); + iommu_register_group(pci->table_group, + pci_domain_nr(pci->phb->bus), 0); + pr_debug(" created table: %p\n", pci->table_group); + } else { + pr_debug(" found DMA window, table: %p\n", pci->table_group); + } + + set_iommu_table_base(&dev->dev, pci->table_group->tables[0]); + iommu_add_device(pci->table_group, &dev->dev); +} + +static bool iommu_bypass_supported_pSeriesLP(struct pci_dev *pdev, u64 dma_mask) +{ + struct device_node *dn = pci_device_to_OF_node(pdev), *pdn; + + /* only attempt to use a new window if 64-bit DMA is requested */ + if (dma_mask < DMA_BIT_MASK(64)) + return false; + + dev_dbg(&pdev->dev, "node is %pOF\n", dn); + + /* + * the device tree might contain the dma-window properties + * per-device and not necessarily for the bus. So we need to + * search upwards in the tree until we either hit a dma-window + * property, OR find a parent with a table already allocated. + */ + pdn = pci_dma_find(dn, NULL); + if (pdn && PCI_DN(pdn)) + return enable_ddw(pdev, pdn); + + return false; +} + +static int iommu_mem_notifier(struct notifier_block *nb, unsigned long action, + void *data) +{ + struct dma_win *window; + struct memory_notify *arg = data; + int ret = 0; + + switch (action) { + case MEM_GOING_ONLINE: + spin_lock(&dma_win_list_lock); + list_for_each_entry(window, &dma_win_list, list) { + if (window->direct) { + ret |= tce_setrange_multi_pSeriesLP(arg->start_pfn, + arg->nr_pages, window->prop); + } + /* XXX log error */ + } + spin_unlock(&dma_win_list_lock); + break; + case MEM_CANCEL_ONLINE: + case MEM_OFFLINE: + spin_lock(&dma_win_list_lock); + list_for_each_entry(window, &dma_win_list, list) { + if (window->direct) { + ret |= tce_clearrange_multi_pSeriesLP(arg->start_pfn, + arg->nr_pages, window->prop); + } + /* XXX log error */ + } + spin_unlock(&dma_win_list_lock); + break; + default: + break; + } + if (ret && action != MEM_CANCEL_ONLINE) + return NOTIFY_BAD; + + return NOTIFY_OK; +} + +static struct notifier_block iommu_mem_nb = { + .notifier_call = iommu_mem_notifier, +}; + +static int iommu_reconfig_notifier(struct notifier_block *nb, unsigned long action, void *data) +{ + int err = NOTIFY_OK; + struct of_reconfig_data *rd = data; + struct device_node *np = rd->dn; + struct pci_dn *pci = PCI_DN(np); + struct dma_win *window; + + switch (action) { + case OF_RECONFIG_DETACH_NODE: + /* + * Removing the property will invoke the reconfig + * notifier again, which causes dead-lock on the + * read-write semaphore of the notifier chain. So + * we have to remove the property when releasing + * the device node. + */ + if (remove_ddw(np, false, DIRECT64_PROPNAME)) + remove_ddw(np, false, DMA64_PROPNAME); + + if (pci && pci->table_group) + iommu_pseries_free_group(pci->table_group, + np->full_name); + + spin_lock(&dma_win_list_lock); + list_for_each_entry(window, &dma_win_list, list) { + if (window->device == np) { + list_del(&window->list); + kfree(window); + break; + } + } + spin_unlock(&dma_win_list_lock); + break; + default: + err = NOTIFY_DONE; + break; + } + return err; +} + +static struct notifier_block iommu_reconfig_nb = { + .notifier_call = iommu_reconfig_notifier, +}; + +/* These are called very early. */ +void __init iommu_init_early_pSeries(void) +{ + if (of_chosen && of_get_property(of_chosen, "linux,iommu-off", NULL)) + return; + + if (firmware_has_feature(FW_FEATURE_LPAR)) { + pseries_pci_controller_ops.dma_bus_setup = pci_dma_bus_setup_pSeriesLP; + pseries_pci_controller_ops.dma_dev_setup = pci_dma_dev_setup_pSeriesLP; + if (!disable_ddw) + pseries_pci_controller_ops.iommu_bypass_supported = + iommu_bypass_supported_pSeriesLP; + } else { + pseries_pci_controller_ops.dma_bus_setup = pci_dma_bus_setup_pSeries; + pseries_pci_controller_ops.dma_dev_setup = pci_dma_dev_setup_pSeries; + } + + + of_reconfig_notifier_register(&iommu_reconfig_nb); + register_memory_notifier(&iommu_mem_nb); + + set_pci_dma_ops(&dma_iommu_ops); +} + +static int __init disable_multitce(char *str) +{ + if (strcmp(str, "off") == 0 && + firmware_has_feature(FW_FEATURE_LPAR) && + (firmware_has_feature(FW_FEATURE_PUT_TCE_IND) || + firmware_has_feature(FW_FEATURE_STUFF_TCE))) { + printk(KERN_INFO "Disabling MULTITCE firmware feature\n"); + powerpc_firmware_features &= + ~(FW_FEATURE_PUT_TCE_IND | FW_FEATURE_STUFF_TCE); + } + return 1; +} + +__setup("multitce=", disable_multitce); + +#ifdef CONFIG_SPAPR_TCE_IOMMU +struct iommu_group *pSeries_pci_device_group(struct pci_controller *hose, + struct pci_dev *pdev) +{ + struct device_node *pdn, *dn = pdev->dev.of_node; + struct iommu_group *grp; + struct pci_dn *pci; + + pdn = pci_dma_find(dn, NULL); + if (!pdn || !PCI_DN(pdn)) + return ERR_PTR(-ENODEV); + + pci = PCI_DN(pdn); + if (!pci->table_group) + return ERR_PTR(-ENODEV); + + grp = pci->table_group->group; + if (!grp) + return ERR_PTR(-ENODEV); + + return iommu_group_ref_get(grp); +} +#endif diff --git a/arch/powerpc/platforms/pseries/kexec.c b/arch/powerpc/platforms/pseries/kexec.c new file mode 100644 index 000000000..096d09ed8 --- /dev/null +++ b/arch/powerpc/platforms/pseries/kexec.c @@ -0,0 +1,71 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright 2006 Michael Ellerman, IBM Corporation + */ + +#include <linux/kernel.h> +#include <linux/interrupt.h> + +#include <asm/setup.h> +#include <asm/page.h> +#include <asm/firmware.h> +#include <asm/kexec.h> +#include <asm/xics.h> +#include <asm/xive.h> +#include <asm/smp.h> +#include <asm/plpar_wrappers.h> + +#include "pseries.h" + +void pseries_kexec_cpu_down(int crash_shutdown, int secondary) +{ + /* + * Don't risk a hypervisor call if we're crashing + * XXX: Why? The hypervisor is not crashing. It might be better + * to at least attempt unregister to avoid the hypervisor stepping + * on our memory. + */ + if (firmware_has_feature(FW_FEATURE_SPLPAR) && !crash_shutdown) { + int ret; + int cpu = smp_processor_id(); + int hwcpu = hard_smp_processor_id(); + + if (get_lppaca()->dtl_enable_mask) { + ret = unregister_dtl(hwcpu); + if (ret) { + pr_err("WARNING: DTL deregistration for cpu " + "%d (hw %d) failed with %d\n", + cpu, hwcpu, ret); + } + } + + ret = unregister_slb_shadow(hwcpu); + if (ret) { + pr_err("WARNING: SLB shadow buffer deregistration " + "for cpu %d (hw %d) failed with %d\n", + cpu, hwcpu, ret); + } + + ret = unregister_vpa(hwcpu); + if (ret) { + pr_err("WARNING: VPA deregistration for cpu %d " + "(hw %d) failed with %d\n", cpu, hwcpu, ret); + } + } + + if (xive_enabled()) { + xive_teardown_cpu(); + + if (!secondary) + xive_shutdown(); + } else + xics_kexec_teardown_cpu(secondary); +} + +void pseries_machine_kexec(struct kimage *image) +{ + if (firmware_has_feature(FW_FEATURE_SET_MODE)) + pseries_disable_reloc_on_exc(); + + default_machine_kexec(image); +} diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c new file mode 100644 index 000000000..d4d6de062 --- /dev/null +++ b/arch/powerpc/platforms/pseries/lpar.c @@ -0,0 +1,2026 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * pSeries_lpar.c + * Copyright (C) 2001 Todd Inglett, IBM Corporation + * + * pSeries LPAR support. + */ + +/* Enables debugging of low-level hash table routines - careful! */ +#undef DEBUG +#define pr_fmt(fmt) "lpar: " fmt + +#include <linux/kernel.h> +#include <linux/dma-mapping.h> +#include <linux/console.h> +#include <linux/export.h> +#include <linux/jump_label.h> +#include <linux/delay.h> +#include <linux/stop_machine.h> +#include <linux/spinlock.h> +#include <linux/cpuhotplug.h> +#include <linux/workqueue.h> +#include <linux/proc_fs.h> +#include <linux/pgtable.h> +#include <linux/debugfs.h> + +#include <asm/processor.h> +#include <asm/mmu.h> +#include <asm/page.h> +#include <asm/setup.h> +#include <asm/mmu_context.h> +#include <asm/iommu.h> +#include <asm/tlb.h> +#include <asm/cputable.h> +#include <asm/papr-sysparm.h> +#include <asm/udbg.h> +#include <asm/smp.h> +#include <asm/trace.h> +#include <asm/firmware.h> +#include <asm/plpar_wrappers.h> +#include <asm/kexec.h> +#include <asm/fadump.h> +#include <asm/dtl.h> +#include <asm/vphn.h> + +#include "pseries.h" + +/* Flag bits for H_BULK_REMOVE */ +#define HBR_REQUEST 0x4000000000000000UL +#define HBR_RESPONSE 0x8000000000000000UL +#define HBR_END 0xc000000000000000UL +#define HBR_AVPN 0x0200000000000000UL +#define HBR_ANDCOND 0x0100000000000000UL + + +/* in hvCall.S */ +EXPORT_SYMBOL(plpar_hcall); +EXPORT_SYMBOL(plpar_hcall9); +EXPORT_SYMBOL(plpar_hcall_norets); + +#ifdef CONFIG_PPC_64S_HASH_MMU +/* + * H_BLOCK_REMOVE supported block size for this page size in segment who's base + * page size is that page size. + * + * The first index is the segment base page size, the second one is the actual + * page size. + */ +static int hblkrm_size[MMU_PAGE_COUNT][MMU_PAGE_COUNT] __ro_after_init; +#endif + +/* + * Due to the involved complexity, and that the current hypervisor is only + * returning this value or 0, we are limiting the support of the H_BLOCK_REMOVE + * buffer size to 8 size block. + */ +#define HBLKRM_SUPPORTED_BLOCK_SIZE 8 + +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE +static u8 dtl_mask = DTL_LOG_PREEMPT; +#else +static u8 dtl_mask; +#endif + +void alloc_dtl_buffers(unsigned long *time_limit) +{ + int cpu; + struct paca_struct *pp; + struct dtl_entry *dtl; + + for_each_possible_cpu(cpu) { + pp = paca_ptrs[cpu]; + if (pp->dispatch_log) + continue; + dtl = kmem_cache_alloc(dtl_cache, GFP_KERNEL); + if (!dtl) { + pr_warn("Failed to allocate dispatch trace log for cpu %d\n", + cpu); +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE + pr_warn("Stolen time statistics will be unreliable\n"); +#endif + break; + } + + pp->dtl_ridx = 0; + pp->dispatch_log = dtl; + pp->dispatch_log_end = dtl + N_DISPATCH_LOG; + pp->dtl_curr = dtl; + + if (time_limit && time_after(jiffies, *time_limit)) { + cond_resched(); + *time_limit = jiffies + HZ; + } + } +} + +void register_dtl_buffer(int cpu) +{ + long ret; + struct paca_struct *pp; + struct dtl_entry *dtl; + int hwcpu = get_hard_smp_processor_id(cpu); + + pp = paca_ptrs[cpu]; + dtl = pp->dispatch_log; + if (dtl && dtl_mask) { + pp->dtl_ridx = 0; + pp->dtl_curr = dtl; + lppaca_of(cpu).dtl_idx = 0; + + /* hypervisor reads buffer length from this field */ + dtl->enqueue_to_dispatch_time = cpu_to_be32(DISPATCH_LOG_BYTES); + ret = register_dtl(hwcpu, __pa(dtl)); + if (ret) + pr_err("WARNING: DTL registration of cpu %d (hw %d) failed with %ld\n", + cpu, hwcpu, ret); + + lppaca_of(cpu).dtl_enable_mask = dtl_mask; + } +} + +#ifdef CONFIG_PPC_SPLPAR +struct dtl_worker { + struct delayed_work work; + int cpu; +}; + +struct vcpu_dispatch_data { + int last_disp_cpu; + + int total_disp; + + int same_cpu_disp; + int same_chip_disp; + int diff_chip_disp; + int far_chip_disp; + + int numa_home_disp; + int numa_remote_disp; + int numa_far_disp; +}; + +/* + * This represents the number of cpus in the hypervisor. Since there is no + * architected way to discover the number of processors in the host, we + * provision for dealing with NR_CPUS. This is currently 2048 by default, and + * is sufficient for our purposes. This will need to be tweaked if + * CONFIG_NR_CPUS is changed. + */ +#define NR_CPUS_H NR_CPUS + +DEFINE_RWLOCK(dtl_access_lock); +static DEFINE_PER_CPU(struct vcpu_dispatch_data, vcpu_disp_data); +static DEFINE_PER_CPU(u64, dtl_entry_ridx); +static DEFINE_PER_CPU(struct dtl_worker, dtl_workers); +static enum cpuhp_state dtl_worker_state; +static DEFINE_MUTEX(dtl_enable_mutex); +static int vcpudispatch_stats_on __read_mostly; +static int vcpudispatch_stats_freq = 50; +static __be32 *vcpu_associativity, *pcpu_associativity; + + +static void free_dtl_buffers(unsigned long *time_limit) +{ +#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE + int cpu; + struct paca_struct *pp; + + for_each_possible_cpu(cpu) { + pp = paca_ptrs[cpu]; + if (!pp->dispatch_log) + continue; + kmem_cache_free(dtl_cache, pp->dispatch_log); + pp->dtl_ridx = 0; + pp->dispatch_log = 0; + pp->dispatch_log_end = 0; + pp->dtl_curr = 0; + + if (time_limit && time_after(jiffies, *time_limit)) { + cond_resched(); + *time_limit = jiffies + HZ; + } + } +#endif +} + +static int init_cpu_associativity(void) +{ + vcpu_associativity = kcalloc(num_possible_cpus() / threads_per_core, + VPHN_ASSOC_BUFSIZE * sizeof(__be32), GFP_KERNEL); + pcpu_associativity = kcalloc(NR_CPUS_H / threads_per_core, + VPHN_ASSOC_BUFSIZE * sizeof(__be32), GFP_KERNEL); + + if (!vcpu_associativity || !pcpu_associativity) { + pr_err("error allocating memory for associativity information\n"); + return -ENOMEM; + } + + return 0; +} + +static void destroy_cpu_associativity(void) +{ + kfree(vcpu_associativity); + kfree(pcpu_associativity); + vcpu_associativity = pcpu_associativity = 0; +} + +static __be32 *__get_cpu_associativity(int cpu, __be32 *cpu_assoc, int flag) +{ + __be32 *assoc; + int rc = 0; + + assoc = &cpu_assoc[(int)(cpu / threads_per_core) * VPHN_ASSOC_BUFSIZE]; + if (!assoc[0]) { + rc = hcall_vphn(cpu, flag, &assoc[0]); + if (rc) + return NULL; + } + + return assoc; +} + +static __be32 *get_pcpu_associativity(int cpu) +{ + return __get_cpu_associativity(cpu, pcpu_associativity, VPHN_FLAG_PCPU); +} + +static __be32 *get_vcpu_associativity(int cpu) +{ + return __get_cpu_associativity(cpu, vcpu_associativity, VPHN_FLAG_VCPU); +} + +static int cpu_relative_dispatch_distance(int last_disp_cpu, int cur_disp_cpu) +{ + __be32 *last_disp_cpu_assoc, *cur_disp_cpu_assoc; + + if (last_disp_cpu >= NR_CPUS_H || cur_disp_cpu >= NR_CPUS_H) + return -EINVAL; + + last_disp_cpu_assoc = get_pcpu_associativity(last_disp_cpu); + cur_disp_cpu_assoc = get_pcpu_associativity(cur_disp_cpu); + + if (!last_disp_cpu_assoc || !cur_disp_cpu_assoc) + return -EIO; + + return cpu_relative_distance(last_disp_cpu_assoc, cur_disp_cpu_assoc); +} + +static int cpu_home_node_dispatch_distance(int disp_cpu) +{ + __be32 *disp_cpu_assoc, *vcpu_assoc; + int vcpu_id = smp_processor_id(); + + if (disp_cpu >= NR_CPUS_H) { + pr_debug_ratelimited("vcpu dispatch cpu %d > %d\n", + disp_cpu, NR_CPUS_H); + return -EINVAL; + } + + disp_cpu_assoc = get_pcpu_associativity(disp_cpu); + vcpu_assoc = get_vcpu_associativity(vcpu_id); + + if (!disp_cpu_assoc || !vcpu_assoc) + return -EIO; + + return cpu_relative_distance(disp_cpu_assoc, vcpu_assoc); +} + +static void update_vcpu_disp_stat(int disp_cpu) +{ + struct vcpu_dispatch_data *disp; + int distance; + + disp = this_cpu_ptr(&vcpu_disp_data); + if (disp->last_disp_cpu == -1) { + disp->last_disp_cpu = disp_cpu; + return; + } + + disp->total_disp++; + + if (disp->last_disp_cpu == disp_cpu || + (cpu_first_thread_sibling(disp->last_disp_cpu) == + cpu_first_thread_sibling(disp_cpu))) + disp->same_cpu_disp++; + else { + distance = cpu_relative_dispatch_distance(disp->last_disp_cpu, + disp_cpu); + if (distance < 0) + pr_debug_ratelimited("vcpudispatch_stats: cpu %d: error determining associativity\n", + smp_processor_id()); + else { + switch (distance) { + case 0: + disp->same_chip_disp++; + break; + case 1: + disp->diff_chip_disp++; + break; + case 2: + disp->far_chip_disp++; + break; + default: + pr_debug_ratelimited("vcpudispatch_stats: cpu %d (%d -> %d): unexpected relative dispatch distance %d\n", + smp_processor_id(), + disp->last_disp_cpu, + disp_cpu, + distance); + } + } + } + + distance = cpu_home_node_dispatch_distance(disp_cpu); + if (distance < 0) + pr_debug_ratelimited("vcpudispatch_stats: cpu %d: error determining associativity\n", + smp_processor_id()); + else { + switch (distance) { + case 0: + disp->numa_home_disp++; + break; + case 1: + disp->numa_remote_disp++; + break; + case 2: + disp->numa_far_disp++; + break; + default: + pr_debug_ratelimited("vcpudispatch_stats: cpu %d on %d: unexpected numa dispatch distance %d\n", + smp_processor_id(), + disp_cpu, + distance); + } + } + + disp->last_disp_cpu = disp_cpu; +} + +static void process_dtl_buffer(struct work_struct *work) +{ + struct dtl_entry dtle; + u64 i = __this_cpu_read(dtl_entry_ridx); + struct dtl_entry *dtl = local_paca->dispatch_log + (i % N_DISPATCH_LOG); + struct dtl_entry *dtl_end = local_paca->dispatch_log_end; + struct lppaca *vpa = local_paca->lppaca_ptr; + struct dtl_worker *d = container_of(work, struct dtl_worker, work.work); + + if (!local_paca->dispatch_log) + return; + + /* if we have been migrated away, we cancel ourself */ + if (d->cpu != smp_processor_id()) { + pr_debug("vcpudispatch_stats: cpu %d worker migrated -- canceling worker\n", + smp_processor_id()); + return; + } + + if (i == be64_to_cpu(vpa->dtl_idx)) + goto out; + + while (i < be64_to_cpu(vpa->dtl_idx)) { + dtle = *dtl; + barrier(); + if (i + N_DISPATCH_LOG < be64_to_cpu(vpa->dtl_idx)) { + /* buffer has overflowed */ + pr_debug_ratelimited("vcpudispatch_stats: cpu %d lost %lld DTL samples\n", + d->cpu, + be64_to_cpu(vpa->dtl_idx) - N_DISPATCH_LOG - i); + i = be64_to_cpu(vpa->dtl_idx) - N_DISPATCH_LOG; + dtl = local_paca->dispatch_log + (i % N_DISPATCH_LOG); + continue; + } + update_vcpu_disp_stat(be16_to_cpu(dtle.processor_id)); + ++i; + ++dtl; + if (dtl == dtl_end) + dtl = local_paca->dispatch_log; + } + + __this_cpu_write(dtl_entry_ridx, i); + +out: + schedule_delayed_work_on(d->cpu, to_delayed_work(work), + HZ / vcpudispatch_stats_freq); +} + +static int dtl_worker_online(unsigned int cpu) +{ + struct dtl_worker *d = &per_cpu(dtl_workers, cpu); + + memset(d, 0, sizeof(*d)); + INIT_DELAYED_WORK(&d->work, process_dtl_buffer); + d->cpu = cpu; + +#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE + per_cpu(dtl_entry_ridx, cpu) = 0; + register_dtl_buffer(cpu); +#else + per_cpu(dtl_entry_ridx, cpu) = be64_to_cpu(lppaca_of(cpu).dtl_idx); +#endif + + schedule_delayed_work_on(cpu, &d->work, HZ / vcpudispatch_stats_freq); + return 0; +} + +static int dtl_worker_offline(unsigned int cpu) +{ + struct dtl_worker *d = &per_cpu(dtl_workers, cpu); + + cancel_delayed_work_sync(&d->work); + +#ifndef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE + unregister_dtl(get_hard_smp_processor_id(cpu)); +#endif + + return 0; +} + +static void set_global_dtl_mask(u8 mask) +{ + int cpu; + + dtl_mask = mask; + for_each_present_cpu(cpu) + lppaca_of(cpu).dtl_enable_mask = dtl_mask; +} + +static void reset_global_dtl_mask(void) +{ + int cpu; + +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE + dtl_mask = DTL_LOG_PREEMPT; +#else + dtl_mask = 0; +#endif + for_each_present_cpu(cpu) + lppaca_of(cpu).dtl_enable_mask = dtl_mask; +} + +static int dtl_worker_enable(unsigned long *time_limit) +{ + int rc = 0, state; + + if (!write_trylock(&dtl_access_lock)) { + rc = -EBUSY; + goto out; + } + + set_global_dtl_mask(DTL_LOG_ALL); + + /* Setup dtl buffers and register those */ + alloc_dtl_buffers(time_limit); + + state = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "powerpc/dtl:online", + dtl_worker_online, dtl_worker_offline); + if (state < 0) { + pr_err("vcpudispatch_stats: unable to setup workqueue for DTL processing\n"); + free_dtl_buffers(time_limit); + reset_global_dtl_mask(); + write_unlock(&dtl_access_lock); + rc = -EINVAL; + goto out; + } + dtl_worker_state = state; + +out: + return rc; +} + +static void dtl_worker_disable(unsigned long *time_limit) +{ + cpuhp_remove_state(dtl_worker_state); + free_dtl_buffers(time_limit); + reset_global_dtl_mask(); + write_unlock(&dtl_access_lock); +} + +static ssize_t vcpudispatch_stats_write(struct file *file, const char __user *p, + size_t count, loff_t *ppos) +{ + unsigned long time_limit = jiffies + HZ; + struct vcpu_dispatch_data *disp; + int rc, cmd, cpu; + char buf[16]; + + if (count > 15) + return -EINVAL; + + if (copy_from_user(buf, p, count)) + return -EFAULT; + + buf[count] = 0; + rc = kstrtoint(buf, 0, &cmd); + if (rc || cmd < 0 || cmd > 1) { + pr_err("vcpudispatch_stats: please use 0 to disable or 1 to enable dispatch statistics\n"); + return rc ? rc : -EINVAL; + } + + mutex_lock(&dtl_enable_mutex); + + if ((cmd == 0 && !vcpudispatch_stats_on) || + (cmd == 1 && vcpudispatch_stats_on)) + goto out; + + if (cmd) { + rc = init_cpu_associativity(); + if (rc) { + destroy_cpu_associativity(); + goto out; + } + + for_each_possible_cpu(cpu) { + disp = per_cpu_ptr(&vcpu_disp_data, cpu); + memset(disp, 0, sizeof(*disp)); + disp->last_disp_cpu = -1; + } + + rc = dtl_worker_enable(&time_limit); + if (rc) { + destroy_cpu_associativity(); + goto out; + } + } else { + dtl_worker_disable(&time_limit); + destroy_cpu_associativity(); + } + + vcpudispatch_stats_on = cmd; + +out: + mutex_unlock(&dtl_enable_mutex); + if (rc) + return rc; + return count; +} + +static int vcpudispatch_stats_display(struct seq_file *p, void *v) +{ + int cpu; + struct vcpu_dispatch_data *disp; + + if (!vcpudispatch_stats_on) { + seq_puts(p, "off\n"); + return 0; + } + + for_each_online_cpu(cpu) { + disp = per_cpu_ptr(&vcpu_disp_data, cpu); + seq_printf(p, "cpu%d", cpu); + seq_put_decimal_ull(p, " ", disp->total_disp); + seq_put_decimal_ull(p, " ", disp->same_cpu_disp); + seq_put_decimal_ull(p, " ", disp->same_chip_disp); + seq_put_decimal_ull(p, " ", disp->diff_chip_disp); + seq_put_decimal_ull(p, " ", disp->far_chip_disp); + seq_put_decimal_ull(p, " ", disp->numa_home_disp); + seq_put_decimal_ull(p, " ", disp->numa_remote_disp); + seq_put_decimal_ull(p, " ", disp->numa_far_disp); + seq_puts(p, "\n"); + } + + return 0; +} + +static int vcpudispatch_stats_open(struct inode *inode, struct file *file) +{ + return single_open(file, vcpudispatch_stats_display, NULL); +} + +static const struct proc_ops vcpudispatch_stats_proc_ops = { + .proc_open = vcpudispatch_stats_open, + .proc_read = seq_read, + .proc_write = vcpudispatch_stats_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static ssize_t vcpudispatch_stats_freq_write(struct file *file, + const char __user *p, size_t count, loff_t *ppos) +{ + int rc, freq; + char buf[16]; + + if (count > 15) + return -EINVAL; + + if (copy_from_user(buf, p, count)) + return -EFAULT; + + buf[count] = 0; + rc = kstrtoint(buf, 0, &freq); + if (rc || freq < 1 || freq > HZ) { + pr_err("vcpudispatch_stats_freq: please specify a frequency between 1 and %d\n", + HZ); + return rc ? rc : -EINVAL; + } + + vcpudispatch_stats_freq = freq; + + return count; +} + +static int vcpudispatch_stats_freq_display(struct seq_file *p, void *v) +{ + seq_printf(p, "%d\n", vcpudispatch_stats_freq); + return 0; +} + +static int vcpudispatch_stats_freq_open(struct inode *inode, struct file *file) +{ + return single_open(file, vcpudispatch_stats_freq_display, NULL); +} + +static const struct proc_ops vcpudispatch_stats_freq_proc_ops = { + .proc_open = vcpudispatch_stats_freq_open, + .proc_read = seq_read, + .proc_write = vcpudispatch_stats_freq_write, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static int __init vcpudispatch_stats_procfs_init(void) +{ + if (!lppaca_shared_proc()) + return 0; + + if (!proc_create("powerpc/vcpudispatch_stats", 0600, NULL, + &vcpudispatch_stats_proc_ops)) + pr_err("vcpudispatch_stats: error creating procfs file\n"); + else if (!proc_create("powerpc/vcpudispatch_stats_freq", 0600, NULL, + &vcpudispatch_stats_freq_proc_ops)) + pr_err("vcpudispatch_stats_freq: error creating procfs file\n"); + + return 0; +} + +machine_device_initcall(pseries, vcpudispatch_stats_procfs_init); + +#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING +u64 pseries_paravirt_steal_clock(int cpu) +{ + struct lppaca *lppaca = &lppaca_of(cpu); + + return be64_to_cpu(READ_ONCE(lppaca->enqueue_dispatch_tb)) + + be64_to_cpu(READ_ONCE(lppaca->ready_enqueue_tb)); +} +#endif + +#endif /* CONFIG_PPC_SPLPAR */ + +void vpa_init(int cpu) +{ + int hwcpu = get_hard_smp_processor_id(cpu); + unsigned long addr; + long ret; + + /* + * The spec says it "may be problematic" if CPU x registers the VPA of + * CPU y. We should never do that, but wail if we ever do. + */ + WARN_ON(cpu != smp_processor_id()); + + if (cpu_has_feature(CPU_FTR_ALTIVEC)) + lppaca_of(cpu).vmxregs_in_use = 1; + + if (cpu_has_feature(CPU_FTR_ARCH_207S)) + lppaca_of(cpu).ebb_regs_in_use = 1; + + addr = __pa(&lppaca_of(cpu)); + ret = register_vpa(hwcpu, addr); + + if (ret) { + pr_err("WARNING: VPA registration for cpu %d (hw %d) of area " + "%lx failed with %ld\n", cpu, hwcpu, addr, ret); + return; + } + +#ifdef CONFIG_PPC_64S_HASH_MMU + /* + * PAPR says this feature is SLB-Buffer but firmware never + * reports that. All SPLPAR support SLB shadow buffer. + */ + if (!radix_enabled() && firmware_has_feature(FW_FEATURE_SPLPAR)) { + addr = __pa(paca_ptrs[cpu]->slb_shadow_ptr); + ret = register_slb_shadow(hwcpu, addr); + if (ret) + pr_err("WARNING: SLB shadow buffer registration for " + "cpu %d (hw %d) of area %lx failed with %ld\n", + cpu, hwcpu, addr, ret); + } +#endif /* CONFIG_PPC_64S_HASH_MMU */ + + /* + * Register dispatch trace log, if one has been allocated. + */ + register_dtl_buffer(cpu); +} + +#ifdef CONFIG_PPC_BOOK3S_64 + +static int __init pseries_lpar_register_process_table(unsigned long base, + unsigned long page_size, unsigned long table_size) +{ + long rc; + unsigned long flags = 0; + + if (table_size) + flags |= PROC_TABLE_NEW; + if (radix_enabled()) { + flags |= PROC_TABLE_RADIX; + if (mmu_has_feature(MMU_FTR_GTSE)) + flags |= PROC_TABLE_GTSE; + } else + flags |= PROC_TABLE_HPT_SLB; + for (;;) { + rc = plpar_hcall_norets(H_REGISTER_PROC_TBL, flags, base, + page_size, table_size); + if (!H_IS_LONG_BUSY(rc)) + break; + mdelay(get_longbusy_msecs(rc)); + } + if (rc != H_SUCCESS) { + pr_err("Failed to register process table (rc=%ld)\n", rc); + BUG(); + } + return rc; +} + +#ifdef CONFIG_PPC_64S_HASH_MMU + +static long pSeries_lpar_hpte_insert(unsigned long hpte_group, + unsigned long vpn, unsigned long pa, + unsigned long rflags, unsigned long vflags, + int psize, int apsize, int ssize) +{ + unsigned long lpar_rc; + unsigned long flags; + unsigned long slot; + unsigned long hpte_v, hpte_r; + + if (!(vflags & HPTE_V_BOLTED)) + pr_devel("hpte_insert(group=%lx, vpn=%016lx, " + "pa=%016lx, rflags=%lx, vflags=%lx, psize=%d)\n", + hpte_group, vpn, pa, rflags, vflags, psize); + + hpte_v = hpte_encode_v(vpn, psize, apsize, ssize) | vflags | HPTE_V_VALID; + hpte_r = hpte_encode_r(pa, psize, apsize) | rflags; + + if (!(vflags & HPTE_V_BOLTED)) + pr_devel(" hpte_v=%016lx, hpte_r=%016lx\n", hpte_v, hpte_r); + + /* Now fill in the actual HPTE */ + /* Set CEC cookie to 0 */ + /* Zero page = 0 */ + /* I-cache Invalidate = 0 */ + /* I-cache synchronize = 0 */ + /* Exact = 0 */ + flags = 0; + + if (firmware_has_feature(FW_FEATURE_XCMO) && !(hpte_r & HPTE_R_N)) + flags |= H_COALESCE_CAND; + + lpar_rc = plpar_pte_enter(flags, hpte_group, hpte_v, hpte_r, &slot); + if (unlikely(lpar_rc == H_PTEG_FULL)) { + pr_devel("Hash table group is full\n"); + return -1; + } + + /* + * Since we try and ioremap PHBs we don't own, the pte insert + * will fail. However we must catch the failure in hash_page + * or we will loop forever, so return -2 in this case. + */ + if (unlikely(lpar_rc != H_SUCCESS)) { + pr_err("Failed hash pte insert with error %ld\n", lpar_rc); + return -2; + } + if (!(vflags & HPTE_V_BOLTED)) + pr_devel(" -> slot: %lu\n", slot & 7); + + /* Because of iSeries, we have to pass down the secondary + * bucket bit here as well + */ + return (slot & 7) | (!!(vflags & HPTE_V_SECONDARY) << 3); +} + +static DEFINE_SPINLOCK(pSeries_lpar_tlbie_lock); + +static long pSeries_lpar_hpte_remove(unsigned long hpte_group) +{ + unsigned long slot_offset; + unsigned long lpar_rc; + int i; + unsigned long dummy1, dummy2; + + /* pick a random slot to start at */ + slot_offset = mftb() & 0x7; + + for (i = 0; i < HPTES_PER_GROUP; i++) { + + /* don't remove a bolted entry */ + lpar_rc = plpar_pte_remove(H_ANDCOND, hpte_group + slot_offset, + HPTE_V_BOLTED, &dummy1, &dummy2); + if (lpar_rc == H_SUCCESS) + return i; + + /* + * The test for adjunct partition is performed before the + * ANDCOND test. H_RESOURCE may be returned, so we need to + * check for that as well. + */ + BUG_ON(lpar_rc != H_NOT_FOUND && lpar_rc != H_RESOURCE); + + slot_offset++; + slot_offset &= 0x7; + } + + return -1; +} + +/* Called during kexec sequence with MMU off */ +static notrace void manual_hpte_clear_all(void) +{ + unsigned long size_bytes = 1UL << ppc64_pft_size; + unsigned long hpte_count = size_bytes >> 4; + struct { + unsigned long pteh; + unsigned long ptel; + } ptes[4]; + long lpar_rc; + unsigned long i, j; + + /* Read in batches of 4, + * invalidate only valid entries not in the VRMA + * hpte_count will be a multiple of 4 + */ + for (i = 0; i < hpte_count; i += 4) { + lpar_rc = plpar_pte_read_4_raw(0, i, (void *)ptes); + if (lpar_rc != H_SUCCESS) { + pr_info("Failed to read hash page table at %ld err %ld\n", + i, lpar_rc); + continue; + } + for (j = 0; j < 4; j++){ + if ((ptes[j].pteh & HPTE_V_VRMA_MASK) == + HPTE_V_VRMA_MASK) + continue; + if (ptes[j].pteh & HPTE_V_VALID) + plpar_pte_remove_raw(0, i + j, 0, + &(ptes[j].pteh), &(ptes[j].ptel)); + } + } +} + +/* Called during kexec sequence with MMU off */ +static notrace int hcall_hpte_clear_all(void) +{ + int rc; + + do { + rc = plpar_hcall_norets(H_CLEAR_HPT); + } while (rc == H_CONTINUE); + + return rc; +} + +/* Called during kexec sequence with MMU off */ +static notrace void pseries_hpte_clear_all(void) +{ + int rc; + + rc = hcall_hpte_clear_all(); + if (rc != H_SUCCESS) + manual_hpte_clear_all(); + +#ifdef __LITTLE_ENDIAN__ + /* + * Reset exceptions to big endian. + * + * FIXME this is a hack for kexec, we need to reset the exception + * endian before starting the new kernel and this is a convenient place + * to do it. + * + * This is also called on boot when a fadump happens. In that case we + * must not change the exception endian mode. + */ + if (firmware_has_feature(FW_FEATURE_SET_MODE) && !is_fadump_active()) + pseries_big_endian_exceptions(); +#endif +} + +/* + * NOTE: for updatepp ops we are fortunate that the linux "newpp" bits and + * the low 3 bits of flags happen to line up. So no transform is needed. + * We can probably optimize here and assume the high bits of newpp are + * already zero. For now I am paranoid. + */ +static long pSeries_lpar_hpte_updatepp(unsigned long slot, + unsigned long newpp, + unsigned long vpn, + int psize, int apsize, + int ssize, unsigned long inv_flags) +{ + unsigned long lpar_rc; + unsigned long flags; + unsigned long want_v; + + want_v = hpte_encode_avpn(vpn, psize, ssize); + + flags = (newpp & (HPTE_R_PP | HPTE_R_N | HPTE_R_KEY_LO)) | H_AVPN; + flags |= (newpp & HPTE_R_KEY_HI) >> 48; + if (mmu_has_feature(MMU_FTR_KERNEL_RO)) + /* Move pp0 into bit 8 (IBM 55) */ + flags |= (newpp & HPTE_R_PP0) >> 55; + + pr_devel(" update: avpnv=%016lx, hash=%016lx, f=%lx, psize: %d ...", + want_v, slot, flags, psize); + + lpar_rc = plpar_pte_protect(flags, slot, want_v); + + if (lpar_rc == H_NOT_FOUND) { + pr_devel("not found !\n"); + return -1; + } + + pr_devel("ok\n"); + + BUG_ON(lpar_rc != H_SUCCESS); + + return 0; +} + +static long __pSeries_lpar_hpte_find(unsigned long want_v, unsigned long hpte_group) +{ + long lpar_rc; + unsigned long i, j; + struct { + unsigned long pteh; + unsigned long ptel; + } ptes[4]; + + for (i = 0; i < HPTES_PER_GROUP; i += 4, hpte_group += 4) { + + lpar_rc = plpar_pte_read_4(0, hpte_group, (void *)ptes); + if (lpar_rc != H_SUCCESS) { + pr_info("Failed to read hash page table at %ld err %ld\n", + hpte_group, lpar_rc); + continue; + } + + for (j = 0; j < 4; j++) { + if (HPTE_V_COMPARE(ptes[j].pteh, want_v) && + (ptes[j].pteh & HPTE_V_VALID)) + return i + j; + } + } + + return -1; +} + +static long pSeries_lpar_hpte_find(unsigned long vpn, int psize, int ssize) +{ + long slot; + unsigned long hash; + unsigned long want_v; + unsigned long hpte_group; + + hash = hpt_hash(vpn, mmu_psize_defs[psize].shift, ssize); + want_v = hpte_encode_avpn(vpn, psize, ssize); + + /* + * We try to keep bolted entries always in primary hash + * But in some case we can find them in secondary too. + */ + hpte_group = (hash & htab_hash_mask) * HPTES_PER_GROUP; + slot = __pSeries_lpar_hpte_find(want_v, hpte_group); + if (slot < 0) { + /* Try in secondary */ + hpte_group = (~hash & htab_hash_mask) * HPTES_PER_GROUP; + slot = __pSeries_lpar_hpte_find(want_v, hpte_group); + if (slot < 0) + return -1; + } + return hpte_group + slot; +} + +static void pSeries_lpar_hpte_updateboltedpp(unsigned long newpp, + unsigned long ea, + int psize, int ssize) +{ + unsigned long vpn; + unsigned long lpar_rc, slot, vsid, flags; + + vsid = get_kernel_vsid(ea, ssize); + vpn = hpt_vpn(ea, vsid, ssize); + + slot = pSeries_lpar_hpte_find(vpn, psize, ssize); + BUG_ON(slot == -1); + + flags = newpp & (HPTE_R_PP | HPTE_R_N); + if (mmu_has_feature(MMU_FTR_KERNEL_RO)) + /* Move pp0 into bit 8 (IBM 55) */ + flags |= (newpp & HPTE_R_PP0) >> 55; + + flags |= ((newpp & HPTE_R_KEY_HI) >> 48) | (newpp & HPTE_R_KEY_LO); + + lpar_rc = plpar_pte_protect(flags, slot, 0); + + BUG_ON(lpar_rc != H_SUCCESS); +} + +static void pSeries_lpar_hpte_invalidate(unsigned long slot, unsigned long vpn, + int psize, int apsize, + int ssize, int local) +{ + unsigned long want_v; + unsigned long lpar_rc; + unsigned long dummy1, dummy2; + + pr_devel(" inval : slot=%lx, vpn=%016lx, psize: %d, local: %d\n", + slot, vpn, psize, local); + + want_v = hpte_encode_avpn(vpn, psize, ssize); + lpar_rc = plpar_pte_remove(H_AVPN, slot, want_v, &dummy1, &dummy2); + if (lpar_rc == H_NOT_FOUND) + return; + + BUG_ON(lpar_rc != H_SUCCESS); +} + + +/* + * As defined in the PAPR's section 14.5.4.1.8 + * The control mask doesn't include the returned reference and change bit from + * the processed PTE. + */ +#define HBLKR_AVPN 0x0100000000000000UL +#define HBLKR_CTRL_MASK 0xf800000000000000UL +#define HBLKR_CTRL_SUCCESS 0x8000000000000000UL +#define HBLKR_CTRL_ERRNOTFOUND 0x8800000000000000UL +#define HBLKR_CTRL_ERRBUSY 0xa000000000000000UL + +/* + * Returned true if we are supporting this block size for the specified segment + * base page size and actual page size. + * + * Currently, we only support 8 size block. + */ +static inline bool is_supported_hlbkrm(int bpsize, int psize) +{ + return (hblkrm_size[bpsize][psize] == HBLKRM_SUPPORTED_BLOCK_SIZE); +} + +/** + * H_BLOCK_REMOVE caller. + * @idx should point to the latest @param entry set with a PTEX. + * If PTE cannot be processed because another CPUs has already locked that + * group, those entries are put back in @param starting at index 1. + * If entries has to be retried and @retry_busy is set to true, these entries + * are retried until success. If @retry_busy is set to false, the returned + * is the number of entries yet to process. + */ +static unsigned long call_block_remove(unsigned long idx, unsigned long *param, + bool retry_busy) +{ + unsigned long i, rc, new_idx; + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE]; + + if (idx < 2) { + pr_warn("Unexpected empty call to H_BLOCK_REMOVE"); + return 0; + } +again: + new_idx = 0; + if (idx > PLPAR_HCALL9_BUFSIZE) { + pr_err("Too many PTEs (%lu) for H_BLOCK_REMOVE", idx); + idx = PLPAR_HCALL9_BUFSIZE; + } else if (idx < PLPAR_HCALL9_BUFSIZE) + param[idx] = HBR_END; + + rc = plpar_hcall9(H_BLOCK_REMOVE, retbuf, + param[0], /* AVA */ + param[1], param[2], param[3], param[4], /* TS0-7 */ + param[5], param[6], param[7], param[8]); + if (rc == H_SUCCESS) + return 0; + + BUG_ON(rc != H_PARTIAL); + + /* Check that the unprocessed entries were 'not found' or 'busy' */ + for (i = 0; i < idx-1; i++) { + unsigned long ctrl = retbuf[i] & HBLKR_CTRL_MASK; + + if (ctrl == HBLKR_CTRL_ERRBUSY) { + param[++new_idx] = param[i+1]; + continue; + } + + BUG_ON(ctrl != HBLKR_CTRL_SUCCESS + && ctrl != HBLKR_CTRL_ERRNOTFOUND); + } + + /* + * If there were entries found busy, retry these entries if requested, + * of if all the entries have to be retried. + */ + if (new_idx && (retry_busy || new_idx == (PLPAR_HCALL9_BUFSIZE-1))) { + idx = new_idx + 1; + goto again; + } + + return new_idx; +} + +#ifdef CONFIG_TRANSPARENT_HUGEPAGE +/* + * Limit iterations holding pSeries_lpar_tlbie_lock to 3. We also need + * to make sure that we avoid bouncing the hypervisor tlbie lock. + */ +#define PPC64_HUGE_HPTE_BATCH 12 + +static void hugepage_block_invalidate(unsigned long *slot, unsigned long *vpn, + int count, int psize, int ssize) +{ + unsigned long param[PLPAR_HCALL9_BUFSIZE]; + unsigned long shift, current_vpgb, vpgb; + int i, pix = 0; + + shift = mmu_psize_defs[psize].shift; + + for (i = 0; i < count; i++) { + /* + * Shifting 3 bits more on the right to get a + * 8 pages aligned virtual addresse. + */ + vpgb = (vpn[i] >> (shift - VPN_SHIFT + 3)); + if (!pix || vpgb != current_vpgb) { + /* + * Need to start a new 8 pages block, flush + * the current one if needed. + */ + if (pix) + (void)call_block_remove(pix, param, true); + current_vpgb = vpgb; + param[0] = hpte_encode_avpn(vpn[i], psize, ssize); + pix = 1; + } + + param[pix++] = HBR_REQUEST | HBLKR_AVPN | slot[i]; + if (pix == PLPAR_HCALL9_BUFSIZE) { + pix = call_block_remove(pix, param, false); + /* + * pix = 0 means that all the entries were + * removed, we can start a new block. + * Otherwise, this means that there are entries + * to retry, and pix points to latest one, so + * we should increment it and try to continue + * the same block. + */ + if (pix) + pix++; + } + } + if (pix) + (void)call_block_remove(pix, param, true); +} + +static void hugepage_bulk_invalidate(unsigned long *slot, unsigned long *vpn, + int count, int psize, int ssize) +{ + unsigned long param[PLPAR_HCALL9_BUFSIZE]; + int i = 0, pix = 0, rc; + + for (i = 0; i < count; i++) { + + if (!firmware_has_feature(FW_FEATURE_BULK_REMOVE)) { + pSeries_lpar_hpte_invalidate(slot[i], vpn[i], psize, 0, + ssize, 0); + } else { + param[pix] = HBR_REQUEST | HBR_AVPN | slot[i]; + param[pix+1] = hpte_encode_avpn(vpn[i], psize, ssize); + pix += 2; + if (pix == 8) { + rc = plpar_hcall9(H_BULK_REMOVE, param, + param[0], param[1], param[2], + param[3], param[4], param[5], + param[6], param[7]); + BUG_ON(rc != H_SUCCESS); + pix = 0; + } + } + } + if (pix) { + param[pix] = HBR_END; + rc = plpar_hcall9(H_BULK_REMOVE, param, param[0], param[1], + param[2], param[3], param[4], param[5], + param[6], param[7]); + BUG_ON(rc != H_SUCCESS); + } +} + +static inline void __pSeries_lpar_hugepage_invalidate(unsigned long *slot, + unsigned long *vpn, + int count, int psize, + int ssize) +{ + unsigned long flags = 0; + int lock_tlbie = !mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE); + + if (lock_tlbie) + spin_lock_irqsave(&pSeries_lpar_tlbie_lock, flags); + + /* Assuming THP size is 16M */ + if (is_supported_hlbkrm(psize, MMU_PAGE_16M)) + hugepage_block_invalidate(slot, vpn, count, psize, ssize); + else + hugepage_bulk_invalidate(slot, vpn, count, psize, ssize); + + if (lock_tlbie) + spin_unlock_irqrestore(&pSeries_lpar_tlbie_lock, flags); +} + +static void pSeries_lpar_hugepage_invalidate(unsigned long vsid, + unsigned long addr, + unsigned char *hpte_slot_array, + int psize, int ssize, int local) +{ + int i, index = 0; + unsigned long s_addr = addr; + unsigned int max_hpte_count, valid; + unsigned long vpn_array[PPC64_HUGE_HPTE_BATCH]; + unsigned long slot_array[PPC64_HUGE_HPTE_BATCH]; + unsigned long shift, hidx, vpn = 0, hash, slot; + + shift = mmu_psize_defs[psize].shift; + max_hpte_count = 1U << (PMD_SHIFT - shift); + + for (i = 0; i < max_hpte_count; i++) { + valid = hpte_valid(hpte_slot_array, i); + if (!valid) + continue; + hidx = hpte_hash_index(hpte_slot_array, i); + + /* get the vpn */ + addr = s_addr + (i * (1ul << shift)); + vpn = hpt_vpn(addr, vsid, ssize); + hash = hpt_hash(vpn, shift, ssize); + if (hidx & _PTEIDX_SECONDARY) + hash = ~hash; + + slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; + slot += hidx & _PTEIDX_GROUP_IX; + + slot_array[index] = slot; + vpn_array[index] = vpn; + if (index == PPC64_HUGE_HPTE_BATCH - 1) { + /* + * Now do a bluk invalidate + */ + __pSeries_lpar_hugepage_invalidate(slot_array, + vpn_array, + PPC64_HUGE_HPTE_BATCH, + psize, ssize); + index = 0; + } else + index++; + } + if (index) + __pSeries_lpar_hugepage_invalidate(slot_array, vpn_array, + index, psize, ssize); +} +#else +static void pSeries_lpar_hugepage_invalidate(unsigned long vsid, + unsigned long addr, + unsigned char *hpte_slot_array, + int psize, int ssize, int local) +{ + WARN(1, "%s called without THP support\n", __func__); +} +#endif + +static int pSeries_lpar_hpte_removebolted(unsigned long ea, + int psize, int ssize) +{ + unsigned long vpn; + unsigned long slot, vsid; + + vsid = get_kernel_vsid(ea, ssize); + vpn = hpt_vpn(ea, vsid, ssize); + + slot = pSeries_lpar_hpte_find(vpn, psize, ssize); + if (slot == -1) + return -ENOENT; + + /* + * lpar doesn't use the passed actual page size + */ + pSeries_lpar_hpte_invalidate(slot, vpn, psize, 0, ssize, 0); + return 0; +} + + +static inline unsigned long compute_slot(real_pte_t pte, + unsigned long vpn, + unsigned long index, + unsigned long shift, + int ssize) +{ + unsigned long slot, hash, hidx; + + hash = hpt_hash(vpn, shift, ssize); + hidx = __rpte_to_hidx(pte, index); + if (hidx & _PTEIDX_SECONDARY) + hash = ~hash; + slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; + slot += hidx & _PTEIDX_GROUP_IX; + return slot; +} + +/** + * The hcall H_BLOCK_REMOVE implies that the virtual pages to processed are + * "all within the same naturally aligned 8 page virtual address block". + */ +static void do_block_remove(unsigned long number, struct ppc64_tlb_batch *batch, + unsigned long *param) +{ + unsigned long vpn; + unsigned long i, pix = 0; + unsigned long index, shift, slot, current_vpgb, vpgb; + real_pte_t pte; + int psize, ssize; + + psize = batch->psize; + ssize = batch->ssize; + + for (i = 0; i < number; i++) { + vpn = batch->vpn[i]; + pte = batch->pte[i]; + pte_iterate_hashed_subpages(pte, psize, vpn, index, shift) { + /* + * Shifting 3 bits more on the right to get a + * 8 pages aligned virtual addresse. + */ + vpgb = (vpn >> (shift - VPN_SHIFT + 3)); + if (!pix || vpgb != current_vpgb) { + /* + * Need to start a new 8 pages block, flush + * the current one if needed. + */ + if (pix) + (void)call_block_remove(pix, param, + true); + current_vpgb = vpgb; + param[0] = hpte_encode_avpn(vpn, psize, + ssize); + pix = 1; + } + + slot = compute_slot(pte, vpn, index, shift, ssize); + param[pix++] = HBR_REQUEST | HBLKR_AVPN | slot; + + if (pix == PLPAR_HCALL9_BUFSIZE) { + pix = call_block_remove(pix, param, false); + /* + * pix = 0 means that all the entries were + * removed, we can start a new block. + * Otherwise, this means that there are entries + * to retry, and pix points to latest one, so + * we should increment it and try to continue + * the same block. + */ + if (pix) + pix++; + } + } pte_iterate_hashed_end(); + } + + if (pix) + (void)call_block_remove(pix, param, true); +} + +/* + * TLB Block Invalidate Characteristics + * + * These characteristics define the size of the block the hcall H_BLOCK_REMOVE + * is able to process for each couple segment base page size, actual page size. + * + * The ibm,get-system-parameter properties is returning a buffer with the + * following layout: + * + * [ 2 bytes size of the RTAS buffer (excluding these 2 bytes) ] + * ----------------- + * TLB Block Invalidate Specifiers: + * [ 1 byte LOG base 2 of the TLB invalidate block size being specified ] + * [ 1 byte Number of page sizes (N) that are supported for the specified + * TLB invalidate block size ] + * [ 1 byte Encoded segment base page size and actual page size + * MSB=0 means 4k segment base page size and actual page size + * MSB=1 the penc value in mmu_psize_def ] + * ... + * ----------------- + * Next TLB Block Invalidate Specifiers... + * ----------------- + * [ 0 ] + */ +static inline void set_hblkrm_bloc_size(int bpsize, int psize, + unsigned int block_size) +{ + if (block_size > hblkrm_size[bpsize][psize]) + hblkrm_size[bpsize][psize] = block_size; +} + +/* + * Decode the Encoded segment base page size and actual page size. + * PAPR specifies: + * - bit 7 is the L bit + * - bits 0-5 are the penc value + * If the L bit is 0, this means 4K segment base page size and actual page size + * otherwise the penc value should be read. + */ +#define HBLKRM_L_MASK 0x80 +#define HBLKRM_PENC_MASK 0x3f +static inline void __init check_lp_set_hblkrm(unsigned int lp, + unsigned int block_size) +{ + unsigned int bpsize, psize; + + /* First, check the L bit, if not set, this means 4K */ + if ((lp & HBLKRM_L_MASK) == 0) { + set_hblkrm_bloc_size(MMU_PAGE_4K, MMU_PAGE_4K, block_size); + return; + } + + lp &= HBLKRM_PENC_MASK; + for (bpsize = 0; bpsize < MMU_PAGE_COUNT; bpsize++) { + struct mmu_psize_def *def = &mmu_psize_defs[bpsize]; + + for (psize = 0; psize < MMU_PAGE_COUNT; psize++) { + if (def->penc[psize] == lp) { + set_hblkrm_bloc_size(bpsize, psize, block_size); + return; + } + } + } +} + +/* + * The size of the TLB Block Invalidate Characteristics is variable. But at the + * maximum it will be the number of possible page sizes *2 + 10 bytes. + * Currently MMU_PAGE_COUNT is 16, which means 42 bytes. Use a cache line size + * (128 bytes) for the buffer to get plenty of space. + */ +#define SPLPAR_TLB_BIC_MAXLENGTH 128 + +void __init pseries_lpar_read_hblkrm_characteristics(void) +{ + static struct papr_sysparm_buf buf __initdata; + int len, idx, bpsize; + + if (!firmware_has_feature(FW_FEATURE_BLOCK_REMOVE)) + return; + + if (papr_sysparm_get(PAPR_SYSPARM_TLB_BLOCK_INVALIDATE_ATTRS, &buf)) + return; + + len = be16_to_cpu(buf.len); + if (len > SPLPAR_TLB_BIC_MAXLENGTH) { + pr_warn("%s too large returned buffer %d", __func__, len); + return; + } + + idx = 0; + while (idx < len) { + u8 block_shift = buf.val[idx++]; + u32 block_size; + unsigned int npsize; + + if (!block_shift) + break; + + block_size = 1 << block_shift; + + for (npsize = buf.val[idx++]; + npsize > 0 && idx < len; npsize--) + check_lp_set_hblkrm((unsigned int)buf.val[idx++], + block_size); + } + + for (bpsize = 0; bpsize < MMU_PAGE_COUNT; bpsize++) + for (idx = 0; idx < MMU_PAGE_COUNT; idx++) + if (hblkrm_size[bpsize][idx]) + pr_info("H_BLOCK_REMOVE supports base psize:%d psize:%d block size:%d", + bpsize, idx, hblkrm_size[bpsize][idx]); +} + +/* + * Take a spinlock around flushes to avoid bouncing the hypervisor tlbie + * lock. + */ +static void pSeries_lpar_flush_hash_range(unsigned long number, int local) +{ + unsigned long vpn; + unsigned long i, pix, rc; + unsigned long flags = 0; + struct ppc64_tlb_batch *batch = this_cpu_ptr(&ppc64_tlb_batch); + int lock_tlbie = !mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE); + unsigned long param[PLPAR_HCALL9_BUFSIZE]; + unsigned long index, shift, slot; + real_pte_t pte; + int psize, ssize; + + if (lock_tlbie) + spin_lock_irqsave(&pSeries_lpar_tlbie_lock, flags); + + if (is_supported_hlbkrm(batch->psize, batch->psize)) { + do_block_remove(number, batch, param); + goto out; + } + + psize = batch->psize; + ssize = batch->ssize; + pix = 0; + for (i = 0; i < number; i++) { + vpn = batch->vpn[i]; + pte = batch->pte[i]; + pte_iterate_hashed_subpages(pte, psize, vpn, index, shift) { + slot = compute_slot(pte, vpn, index, shift, ssize); + if (!firmware_has_feature(FW_FEATURE_BULK_REMOVE)) { + /* + * lpar doesn't use the passed actual page size + */ + pSeries_lpar_hpte_invalidate(slot, vpn, psize, + 0, ssize, local); + } else { + param[pix] = HBR_REQUEST | HBR_AVPN | slot; + param[pix+1] = hpte_encode_avpn(vpn, psize, + ssize); + pix += 2; + if (pix == 8) { + rc = plpar_hcall9(H_BULK_REMOVE, param, + param[0], param[1], param[2], + param[3], param[4], param[5], + param[6], param[7]); + BUG_ON(rc != H_SUCCESS); + pix = 0; + } + } + } pte_iterate_hashed_end(); + } + if (pix) { + param[pix] = HBR_END; + rc = plpar_hcall9(H_BULK_REMOVE, param, param[0], param[1], + param[2], param[3], param[4], param[5], + param[6], param[7]); + BUG_ON(rc != H_SUCCESS); + } + +out: + if (lock_tlbie) + spin_unlock_irqrestore(&pSeries_lpar_tlbie_lock, flags); +} + +static int __init disable_bulk_remove(char *str) +{ + if (strcmp(str, "off") == 0 && + firmware_has_feature(FW_FEATURE_BULK_REMOVE)) { + pr_info("Disabling BULK_REMOVE firmware feature"); + powerpc_firmware_features &= ~FW_FEATURE_BULK_REMOVE; + } + return 1; +} + +__setup("bulk_remove=", disable_bulk_remove); + +#define HPT_RESIZE_TIMEOUT 10000 /* ms */ + +struct hpt_resize_state { + unsigned long shift; + int commit_rc; +}; + +static int pseries_lpar_resize_hpt_commit(void *data) +{ + struct hpt_resize_state *state = data; + + state->commit_rc = plpar_resize_hpt_commit(0, state->shift); + if (state->commit_rc != H_SUCCESS) + return -EIO; + + /* Hypervisor has transitioned the HTAB, update our globals */ + ppc64_pft_size = state->shift; + htab_size_bytes = 1UL << ppc64_pft_size; + htab_hash_mask = (htab_size_bytes >> 7) - 1; + + return 0; +} + +/* + * Must be called in process context. The caller must hold the + * cpus_lock. + */ +static int pseries_lpar_resize_hpt(unsigned long shift) +{ + struct hpt_resize_state state = { + .shift = shift, + .commit_rc = H_FUNCTION, + }; + unsigned int delay, total_delay = 0; + int rc; + ktime_t t0, t1, t2; + + might_sleep(); + + if (!firmware_has_feature(FW_FEATURE_HPT_RESIZE)) + return -ENODEV; + + pr_info("Attempting to resize HPT to shift %lu\n", shift); + + t0 = ktime_get(); + + rc = plpar_resize_hpt_prepare(0, shift); + while (H_IS_LONG_BUSY(rc)) { + delay = get_longbusy_msecs(rc); + total_delay += delay; + if (total_delay > HPT_RESIZE_TIMEOUT) { + /* prepare with shift==0 cancels an in-progress resize */ + rc = plpar_resize_hpt_prepare(0, 0); + if (rc != H_SUCCESS) + pr_warn("Unexpected error %d cancelling timed out HPT resize\n", + rc); + return -ETIMEDOUT; + } + msleep(delay); + rc = plpar_resize_hpt_prepare(0, shift); + } + + switch (rc) { + case H_SUCCESS: + /* Continue on */ + break; + + case H_PARAMETER: + pr_warn("Invalid argument from H_RESIZE_HPT_PREPARE\n"); + return -EINVAL; + case H_RESOURCE: + pr_warn("Operation not permitted from H_RESIZE_HPT_PREPARE\n"); + return -EPERM; + default: + pr_warn("Unexpected error %d from H_RESIZE_HPT_PREPARE\n", rc); + return -EIO; + } + + t1 = ktime_get(); + + rc = stop_machine_cpuslocked(pseries_lpar_resize_hpt_commit, + &state, NULL); + + t2 = ktime_get(); + + if (rc != 0) { + switch (state.commit_rc) { + case H_PTEG_FULL: + return -ENOSPC; + + default: + pr_warn("Unexpected error %d from H_RESIZE_HPT_COMMIT\n", + state.commit_rc); + return -EIO; + }; + } + + pr_info("HPT resize to shift %lu complete (%lld ms / %lld ms)\n", + shift, (long long) ktime_ms_delta(t1, t0), + (long long) ktime_ms_delta(t2, t1)); + + return 0; +} + +void __init hpte_init_pseries(void) +{ + mmu_hash_ops.hpte_invalidate = pSeries_lpar_hpte_invalidate; + mmu_hash_ops.hpte_updatepp = pSeries_lpar_hpte_updatepp; + mmu_hash_ops.hpte_updateboltedpp = pSeries_lpar_hpte_updateboltedpp; + mmu_hash_ops.hpte_insert = pSeries_lpar_hpte_insert; + mmu_hash_ops.hpte_remove = pSeries_lpar_hpte_remove; + mmu_hash_ops.hpte_removebolted = pSeries_lpar_hpte_removebolted; + mmu_hash_ops.flush_hash_range = pSeries_lpar_flush_hash_range; + mmu_hash_ops.hpte_clear_all = pseries_hpte_clear_all; + mmu_hash_ops.hugepage_invalidate = pSeries_lpar_hugepage_invalidate; + + if (firmware_has_feature(FW_FEATURE_HPT_RESIZE)) + mmu_hash_ops.resize_hpt = pseries_lpar_resize_hpt; + + /* + * On POWER9, we need to do a H_REGISTER_PROC_TBL hcall + * to inform the hypervisor that we wish to use the HPT. + */ + if (cpu_has_feature(CPU_FTR_ARCH_300)) + pseries_lpar_register_process_table(0, 0, 0); +} +#endif /* CONFIG_PPC_64S_HASH_MMU */ + +#ifdef CONFIG_PPC_RADIX_MMU +void __init radix_init_pseries(void) +{ + pr_info("Using radix MMU under hypervisor\n"); + + pseries_lpar_register_process_table(__pa(process_tb), + 0, PRTB_SIZE_SHIFT - 12); +} +#endif + +#ifdef CONFIG_PPC_SMLPAR +#define CMO_FREE_HINT_DEFAULT 1 +static int cmo_free_hint_flag = CMO_FREE_HINT_DEFAULT; + +static int __init cmo_free_hint(char *str) +{ + char *parm; + parm = strstrip(str); + + if (strcasecmp(parm, "no") == 0 || strcasecmp(parm, "off") == 0) { + pr_info("%s: CMO free page hinting is not active.\n", __func__); + cmo_free_hint_flag = 0; + return 1; + } + + cmo_free_hint_flag = 1; + pr_info("%s: CMO free page hinting is active.\n", __func__); + + if (strcasecmp(parm, "yes") == 0 || strcasecmp(parm, "on") == 0) + return 1; + + return 0; +} + +__setup("cmo_free_hint=", cmo_free_hint); + +static void pSeries_set_page_state(struct page *page, int order, + unsigned long state) +{ + int i, j; + unsigned long cmo_page_sz, addr; + + cmo_page_sz = cmo_get_page_size(); + addr = __pa((unsigned long)page_address(page)); + + for (i = 0; i < (1 << order); i++, addr += PAGE_SIZE) { + for (j = 0; j < PAGE_SIZE; j += cmo_page_sz) + plpar_hcall_norets(H_PAGE_INIT, state, addr + j, 0); + } +} + +void arch_free_page(struct page *page, int order) +{ + if (radix_enabled()) + return; + if (!cmo_free_hint_flag || !firmware_has_feature(FW_FEATURE_CMO)) + return; + + pSeries_set_page_state(page, order, H_PAGE_SET_UNUSED); +} +EXPORT_SYMBOL(arch_free_page); + +#endif /* CONFIG_PPC_SMLPAR */ +#endif /* CONFIG_PPC_BOOK3S_64 */ + +#ifdef CONFIG_TRACEPOINTS +#ifdef CONFIG_JUMP_LABEL +struct static_key hcall_tracepoint_key = STATIC_KEY_INIT; + +int hcall_tracepoint_regfunc(void) +{ + static_key_slow_inc(&hcall_tracepoint_key); + return 0; +} + +void hcall_tracepoint_unregfunc(void) +{ + static_key_slow_dec(&hcall_tracepoint_key); +} +#else +/* + * We optimise our hcall path by placing hcall_tracepoint_refcount + * directly in the TOC so we can check if the hcall tracepoints are + * enabled via a single load. + */ + +/* NB: reg/unreg are called while guarded with the tracepoints_mutex */ +extern long hcall_tracepoint_refcount; + +int hcall_tracepoint_regfunc(void) +{ + hcall_tracepoint_refcount++; + return 0; +} + +void hcall_tracepoint_unregfunc(void) +{ + hcall_tracepoint_refcount--; +} +#endif + +/* + * Keep track of hcall tracing depth and prevent recursion. Warn if any is + * detected because it may indicate a problem. This will not catch all + * problems with tracing code making hcalls, because the tracing might have + * been invoked from a non-hcall, so the first hcall could recurse into it + * without warning here, but this better than nothing. + * + * Hcalls with specific problems being traced should use the _notrace + * plpar_hcall variants. + */ +static DEFINE_PER_CPU(unsigned int, hcall_trace_depth); + + +notrace void __trace_hcall_entry(unsigned long opcode, unsigned long *args) +{ + unsigned long flags; + unsigned int *depth; + + local_irq_save(flags); + + depth = this_cpu_ptr(&hcall_trace_depth); + + if (WARN_ON_ONCE(*depth)) + goto out; + + (*depth)++; + preempt_disable(); + trace_hcall_entry(opcode, args); + (*depth)--; + +out: + local_irq_restore(flags); +} + +notrace void __trace_hcall_exit(long opcode, long retval, unsigned long *retbuf) +{ + unsigned long flags; + unsigned int *depth; + + local_irq_save(flags); + + depth = this_cpu_ptr(&hcall_trace_depth); + + if (*depth) /* Don't warn again on the way out */ + goto out; + + (*depth)++; + trace_hcall_exit(opcode, retval, retbuf); + preempt_enable(); + (*depth)--; + +out: + local_irq_restore(flags); +} +#endif + +/** + * h_get_mpp + * H_GET_MPP hcall returns info in 7 parms + */ +int h_get_mpp(struct hvcall_mpp_data *mpp_data) +{ + int rc; + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE]; + + rc = plpar_hcall9(H_GET_MPP, retbuf); + + mpp_data->entitled_mem = retbuf[0]; + mpp_data->mapped_mem = retbuf[1]; + + mpp_data->group_num = (retbuf[2] >> 2 * 8) & 0xffff; + mpp_data->pool_num = retbuf[2] & 0xffff; + + mpp_data->mem_weight = (retbuf[3] >> 7 * 8) & 0xff; + mpp_data->unallocated_mem_weight = (retbuf[3] >> 6 * 8) & 0xff; + mpp_data->unallocated_entitlement = retbuf[3] & 0xffffffffffffUL; + + mpp_data->pool_size = retbuf[4]; + mpp_data->loan_request = retbuf[5]; + mpp_data->backing_mem = retbuf[6]; + + return rc; +} +EXPORT_SYMBOL(h_get_mpp); + +int h_get_mpp_x(struct hvcall_mpp_x_data *mpp_x_data) +{ + int rc; + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE] = { 0 }; + + rc = plpar_hcall9(H_GET_MPP_X, retbuf); + + mpp_x_data->coalesced_bytes = retbuf[0]; + mpp_x_data->pool_coalesced_bytes = retbuf[1]; + mpp_x_data->pool_purr_cycles = retbuf[2]; + mpp_x_data->pool_spurr_cycles = retbuf[3]; + + return rc; +} + +#ifdef CONFIG_PPC_64S_HASH_MMU +static unsigned long __init vsid_unscramble(unsigned long vsid, int ssize) +{ + unsigned long protovsid; + unsigned long va_bits = VA_BITS; + unsigned long modinv, vsid_modulus; + unsigned long max_mod_inv, tmp_modinv; + + if (!mmu_has_feature(MMU_FTR_68_BIT_VA)) + va_bits = 65; + + if (ssize == MMU_SEGSIZE_256M) { + modinv = VSID_MULINV_256M; + vsid_modulus = ((1UL << (va_bits - SID_SHIFT)) - 1); + } else { + modinv = VSID_MULINV_1T; + vsid_modulus = ((1UL << (va_bits - SID_SHIFT_1T)) - 1); + } + + /* + * vsid outside our range. + */ + if (vsid >= vsid_modulus) + return 0; + + /* + * If modinv is the modular multiplicate inverse of (x % vsid_modulus) + * and vsid = (protovsid * x) % vsid_modulus, then we say: + * protovsid = (vsid * modinv) % vsid_modulus + */ + + /* Check if (vsid * modinv) overflow (63 bits) */ + max_mod_inv = 0x7fffffffffffffffull / vsid; + if (modinv < max_mod_inv) + return (vsid * modinv) % vsid_modulus; + + tmp_modinv = modinv/max_mod_inv; + modinv %= max_mod_inv; + + protovsid = (((vsid * max_mod_inv) % vsid_modulus) * tmp_modinv) % vsid_modulus; + protovsid = (protovsid + vsid * modinv) % vsid_modulus; + + return protovsid; +} + +static int __init reserve_vrma_context_id(void) +{ + unsigned long protovsid; + + /* + * Reserve context ids which map to reserved virtual addresses. For now + * we only reserve the context id which maps to the VRMA VSID. We ignore + * the addresses in "ibm,adjunct-virtual-addresses" because we don't + * enable adjunct support via the "ibm,client-architecture-support" + * interface. + */ + protovsid = vsid_unscramble(VRMA_VSID, MMU_SEGSIZE_1T); + hash__reserve_context_id(protovsid >> ESID_BITS_1T); + return 0; +} +machine_device_initcall(pseries, reserve_vrma_context_id); +#endif + +#ifdef CONFIG_DEBUG_FS +/* debugfs file interface for vpa data */ +static ssize_t vpa_file_read(struct file *filp, char __user *buf, size_t len, + loff_t *pos) +{ + int cpu = (long)filp->private_data; + struct lppaca *lppaca = &lppaca_of(cpu); + + return simple_read_from_buffer(buf, len, pos, lppaca, + sizeof(struct lppaca)); +} + +static const struct file_operations vpa_fops = { + .open = simple_open, + .read = vpa_file_read, + .llseek = default_llseek, +}; + +static int __init vpa_debugfs_init(void) +{ + char name[16]; + long i; + struct dentry *vpa_dir; + + if (!firmware_has_feature(FW_FEATURE_SPLPAR)) + return 0; + + vpa_dir = debugfs_create_dir("vpa", arch_debugfs_dir); + + /* set up the per-cpu vpa file*/ + for_each_possible_cpu(i) { + sprintf(name, "cpu-%ld", i); + debugfs_create_file(name, 0400, vpa_dir, (void *)i, &vpa_fops); + } + + return 0; +} +machine_arch_initcall(pseries, vpa_debugfs_init); +#endif /* CONFIG_DEBUG_FS */ diff --git a/arch/powerpc/platforms/pseries/lparcfg.c b/arch/powerpc/platforms/pseries/lparcfg.c new file mode 100644 index 000000000..1c151d77e --- /dev/null +++ b/arch/powerpc/platforms/pseries/lparcfg.c @@ -0,0 +1,802 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * PowerPC64 LPAR Configuration Information Driver + * + * Dave Engebretsen engebret@us.ibm.com + * Copyright (c) 2003 Dave Engebretsen + * Will Schmidt willschm@us.ibm.com + * SPLPAR updates, Copyright (c) 2003 Will Schmidt IBM Corporation. + * seq_file updates, Copyright (c) 2004 Will Schmidt IBM Corporation. + * Nathan Lynch nathanl@austin.ibm.com + * Added lparcfg_write, Copyright (C) 2004 Nathan Lynch IBM Corporation. + * + * This driver creates a proc file at /proc/ppc64/lparcfg which contains + * keyword - value pairs that specify the configuration of the partition. + */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/errno.h> +#include <linux/proc_fs.h> +#include <linux/init.h> +#include <asm/papr-sysparm.h> +#include <linux/seq_file.h> +#include <linux/slab.h> +#include <linux/uaccess.h> +#include <linux/hugetlb.h> +#include <asm/lppaca.h> +#include <asm/hvcall.h> +#include <asm/firmware.h> +#include <asm/rtas.h> +#include <asm/time.h> +#include <asm/vdso_datapage.h> +#include <asm/vio.h> +#include <asm/mmu.h> +#include <asm/machdep.h> +#include <asm/drmem.h> + +#include "pseries.h" +#include "vas.h" /* pseries_vas_dlpar_cpu() */ + +/* + * This isn't a module but we expose that to userspace + * via /proc so leave the definitions here + */ +#define MODULE_VERS "1.9" +#define MODULE_NAME "lparcfg" + +/* #define LPARCFG_DEBUG */ + +/* + * Track sum of all purrs across all processors. This is used to further + * calculate usage values by different applications + */ +static void cpu_get_purr(void *arg) +{ + atomic64_t *sum = arg; + + atomic64_add(mfspr(SPRN_PURR), sum); +} + +static unsigned long get_purr(void) +{ + atomic64_t purr = ATOMIC64_INIT(0); + + on_each_cpu(cpu_get_purr, &purr, 1); + + return atomic64_read(&purr); +} + +/* + * Methods used to fetch LPAR data when running on a pSeries platform. + */ + +struct hvcall_ppp_data { + u64 entitlement; + u64 unallocated_entitlement; + u16 group_num; + u16 pool_num; + u8 capped; + u8 weight; + u8 unallocated_weight; + u16 active_procs_in_pool; + u16 active_system_procs; + u16 phys_platform_procs; + u32 max_proc_cap_avail; + u32 entitled_proc_cap_avail; +}; + +/* + * H_GET_PPP hcall returns info in 4 parms. + * entitled_capacity,unallocated_capacity, + * aggregation, resource_capability). + * + * R4 = Entitled Processor Capacity Percentage. + * R5 = Unallocated Processor Capacity Percentage. + * R6 (AABBCCDDEEFFGGHH). + * XXXX - reserved (0) + * XXXX - reserved (0) + * XXXX - Group Number + * XXXX - Pool Number. + * R7 (IIJJKKLLMMNNOOPP). + * XX - reserved. (0) + * XX - bit 0-6 reserved (0). bit 7 is Capped indicator. + * XX - variable processor Capacity Weight + * XX - Unallocated Variable Processor Capacity Weight. + * XXXX - Active processors in Physical Processor Pool. + * XXXX - Processors active on platform. + * R8 (QQQQRRRRRRSSSSSS). if ibm,partition-performance-parameters-level >= 1 + * XXXX - Physical platform procs allocated to virtualization. + * XXXXXX - Max procs capacity % available to the partitions pool. + * XXXXXX - Entitled procs capacity % available to the + * partitions pool. + */ +static unsigned int h_get_ppp(struct hvcall_ppp_data *ppp_data) +{ + unsigned long rc; + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE]; + + rc = plpar_hcall9(H_GET_PPP, retbuf); + + ppp_data->entitlement = retbuf[0]; + ppp_data->unallocated_entitlement = retbuf[1]; + + ppp_data->group_num = (retbuf[2] >> 2 * 8) & 0xffff; + ppp_data->pool_num = retbuf[2] & 0xffff; + + ppp_data->capped = (retbuf[3] >> 6 * 8) & 0x01; + ppp_data->weight = (retbuf[3] >> 5 * 8) & 0xff; + ppp_data->unallocated_weight = (retbuf[3] >> 4 * 8) & 0xff; + ppp_data->active_procs_in_pool = (retbuf[3] >> 2 * 8) & 0xffff; + ppp_data->active_system_procs = retbuf[3] & 0xffff; + + ppp_data->phys_platform_procs = retbuf[4] >> 6 * 8; + ppp_data->max_proc_cap_avail = (retbuf[4] >> 3 * 8) & 0xffffff; + ppp_data->entitled_proc_cap_avail = retbuf[4] & 0xffffff; + + return rc; +} + +static void show_gpci_data(struct seq_file *m) +{ + struct hv_gpci_request_buffer *buf; + unsigned int affinity_score; + long ret; + + buf = kmalloc(sizeof(*buf), GFP_KERNEL); + if (buf == NULL) + return; + + /* + * Show the local LPAR's affinity score. + * + * 0xB1 selects the Affinity_Domain_Info_By_Partition subcall. + * The score is at byte 0xB in the output buffer. + */ + memset(&buf->params, 0, sizeof(buf->params)); + buf->params.counter_request = cpu_to_be32(0xB1); + buf->params.starting_index = cpu_to_be32(-1); /* local LPAR */ + buf->params.counter_info_version_in = 0x5; /* v5+ for score */ + ret = plpar_hcall_norets(H_GET_PERF_COUNTER_INFO, virt_to_phys(buf), + sizeof(*buf)); + if (ret != H_SUCCESS) { + pr_debug("hcall failed: H_GET_PERF_COUNTER_INFO: %ld, %x\n", + ret, be32_to_cpu(buf->params.detail_rc)); + goto out; + } + affinity_score = buf->bytes[0xB]; + seq_printf(m, "partition_affinity_score=%u\n", affinity_score); +out: + kfree(buf); +} + +static unsigned h_pic(unsigned long *pool_idle_time, + unsigned long *num_procs) +{ + unsigned long rc; + unsigned long retbuf[PLPAR_HCALL_BUFSIZE]; + + rc = plpar_hcall(H_PIC, retbuf); + + *pool_idle_time = retbuf[0]; + *num_procs = retbuf[1]; + + return rc; +} + +/* + * parse_ppp_data + * Parse out the data returned from h_get_ppp and h_pic + */ +static void parse_ppp_data(struct seq_file *m) +{ + struct hvcall_ppp_data ppp_data; + struct device_node *root; + const __be32 *perf_level; + int rc; + + rc = h_get_ppp(&ppp_data); + if (rc) + return; + + seq_printf(m, "partition_entitled_capacity=%lld\n", + ppp_data.entitlement); + seq_printf(m, "group=%d\n", ppp_data.group_num); + seq_printf(m, "system_active_processors=%d\n", + ppp_data.active_system_procs); + + /* pool related entries are appropriate for shared configs */ + if (lppaca_shared_proc()) { + unsigned long pool_idle_time, pool_procs; + + seq_printf(m, "pool=%d\n", ppp_data.pool_num); + + /* report pool_capacity in percentage */ + seq_printf(m, "pool_capacity=%d\n", + ppp_data.active_procs_in_pool * 100); + + h_pic(&pool_idle_time, &pool_procs); + seq_printf(m, "pool_idle_time=%ld\n", pool_idle_time); + seq_printf(m, "pool_num_procs=%ld\n", pool_procs); + } + + seq_printf(m, "unallocated_capacity_weight=%d\n", + ppp_data.unallocated_weight); + seq_printf(m, "capacity_weight=%d\n", ppp_data.weight); + seq_printf(m, "capped=%d\n", ppp_data.capped); + seq_printf(m, "unallocated_capacity=%lld\n", + ppp_data.unallocated_entitlement); + + /* The last bits of information returned from h_get_ppp are only + * valid if the ibm,partition-performance-parameters-level + * property is >= 1. + */ + root = of_find_node_by_path("/"); + if (root) { + perf_level = of_get_property(root, + "ibm,partition-performance-parameters-level", + NULL); + if (perf_level && (be32_to_cpup(perf_level) >= 1)) { + seq_printf(m, + "physical_procs_allocated_to_virtualization=%d\n", + ppp_data.phys_platform_procs); + seq_printf(m, "max_proc_capacity_available=%d\n", + ppp_data.max_proc_cap_avail); + seq_printf(m, "entitled_proc_capacity_available=%d\n", + ppp_data.entitled_proc_cap_avail); + } + + of_node_put(root); + } +} + +/** + * parse_mpp_data + * Parse out data returned from h_get_mpp + */ +static void parse_mpp_data(struct seq_file *m) +{ + struct hvcall_mpp_data mpp_data; + int rc; + + rc = h_get_mpp(&mpp_data); + if (rc) + return; + + seq_printf(m, "entitled_memory=%ld\n", mpp_data.entitled_mem); + + if (mpp_data.mapped_mem != -1) + seq_printf(m, "mapped_entitled_memory=%ld\n", + mpp_data.mapped_mem); + + seq_printf(m, "entitled_memory_group_number=%d\n", mpp_data.group_num); + seq_printf(m, "entitled_memory_pool_number=%d\n", mpp_data.pool_num); + + seq_printf(m, "entitled_memory_weight=%d\n", mpp_data.mem_weight); + seq_printf(m, "unallocated_entitled_memory_weight=%d\n", + mpp_data.unallocated_mem_weight); + seq_printf(m, "unallocated_io_mapping_entitlement=%ld\n", + mpp_data.unallocated_entitlement); + + if (mpp_data.pool_size != -1) + seq_printf(m, "entitled_memory_pool_size=%ld bytes\n", + mpp_data.pool_size); + + seq_printf(m, "entitled_memory_loan_request=%ld\n", + mpp_data.loan_request); + + seq_printf(m, "backing_memory=%ld bytes\n", mpp_data.backing_mem); +} + +/** + * parse_mpp_x_data + * Parse out data returned from h_get_mpp_x + */ +static void parse_mpp_x_data(struct seq_file *m) +{ + struct hvcall_mpp_x_data mpp_x_data; + + if (!firmware_has_feature(FW_FEATURE_XCMO)) + return; + if (h_get_mpp_x(&mpp_x_data)) + return; + + seq_printf(m, "coalesced_bytes=%ld\n", mpp_x_data.coalesced_bytes); + + if (mpp_x_data.pool_coalesced_bytes) + seq_printf(m, "pool_coalesced_bytes=%ld\n", + mpp_x_data.pool_coalesced_bytes); + if (mpp_x_data.pool_purr_cycles) + seq_printf(m, "coalesce_pool_purr=%ld\n", mpp_x_data.pool_purr_cycles); + if (mpp_x_data.pool_spurr_cycles) + seq_printf(m, "coalesce_pool_spurr=%ld\n", mpp_x_data.pool_spurr_cycles); +} + +/* + * Read the lpar name using the RTAS ibm,get-system-parameter call. + * + * The name read through this call is updated if changes are made by the end + * user on the hypervisor side. + * + * Some hypervisor (like Qemu) may not provide this value. In that case, a non + * null value is returned. + */ +static int read_rtas_lpar_name(struct seq_file *m) +{ + struct papr_sysparm_buf *buf; + int err; + + buf = papr_sysparm_buf_alloc(); + if (!buf) + return -ENOMEM; + + err = papr_sysparm_get(PAPR_SYSPARM_LPAR_NAME, buf); + if (!err) + seq_printf(m, "partition_name=%s\n", buf->val); + + papr_sysparm_buf_free(buf); + return err; +} + +/* + * Read the LPAR name from the Device Tree. + * + * The value read in the DT is not updated if the end-user is touching the LPAR + * name on the hypervisor side. + */ +static int read_dt_lpar_name(struct seq_file *m) +{ + const char *name; + + if (of_property_read_string(of_root, "ibm,partition-name", &name)) + return -ENOENT; + + seq_printf(m, "partition_name=%s\n", name); + return 0; +} + +static void read_lpar_name(struct seq_file *m) +{ + if (read_rtas_lpar_name(m) && read_dt_lpar_name(m)) + pr_err_once("Error can't get the LPAR name"); +} + +#define SPLPAR_MAXLENGTH 1026*(sizeof(char)) + +/* + * parse_system_parameter_string() + * Retrieve the potential_processors, max_entitled_capacity and friends + * through the get-system-parameter rtas call. Replace keyword strings as + * necessary. + */ +static void parse_system_parameter_string(struct seq_file *m) +{ + struct papr_sysparm_buf *buf; + + buf = papr_sysparm_buf_alloc(); + if (!buf) + return; + + if (papr_sysparm_get(PAPR_SYSPARM_SHARED_PROC_LPAR_ATTRS, buf)) { + goto out_free; + } else { + const char *local_buffer; + int splpar_strlen; + int idx, w_idx; + char *workbuffer = kzalloc(SPLPAR_MAXLENGTH, GFP_KERNEL); + + if (!workbuffer) + goto out_free; + + splpar_strlen = be16_to_cpu(buf->len); + local_buffer = buf->val; + + w_idx = 0; + idx = 0; + while ((*local_buffer) && (idx < splpar_strlen)) { + workbuffer[w_idx++] = local_buffer[idx++]; + if ((local_buffer[idx] == ',') + || (local_buffer[idx] == '\0')) { + workbuffer[w_idx] = '\0'; + if (w_idx) { + /* avoid the empty string */ + seq_printf(m, "%s\n", workbuffer); + } + memset(workbuffer, 0, SPLPAR_MAXLENGTH); + idx++; /* skip the comma */ + w_idx = 0; + } else if (local_buffer[idx] == '=') { + /* code here to replace workbuffer contents + with different keyword strings */ + if (0 == strcmp(workbuffer, "MaxEntCap")) { + strcpy(workbuffer, + "partition_max_entitled_capacity"); + w_idx = strlen(workbuffer); + } + if (0 == strcmp(workbuffer, "MaxPlatProcs")) { + strcpy(workbuffer, + "system_potential_processors"); + w_idx = strlen(workbuffer); + } + } + } + kfree(workbuffer); + local_buffer -= 2; /* back up over strlen value */ + } +out_free: + papr_sysparm_buf_free(buf); +} + +/* Return the number of processors in the system. + * This function reads through the device tree and counts + * the virtual processors, this does not include threads. + */ +static int lparcfg_count_active_processors(void) +{ + struct device_node *cpus_dn; + int count = 0; + + for_each_node_by_type(cpus_dn, "cpu") { +#ifdef LPARCFG_DEBUG + printk(KERN_ERR "cpus_dn %p\n", cpus_dn); +#endif + count++; + } + return count; +} + +static void pseries_cmo_data(struct seq_file *m) +{ + int cpu; + unsigned long cmo_faults = 0; + unsigned long cmo_fault_time = 0; + + seq_printf(m, "cmo_enabled=%d\n", firmware_has_feature(FW_FEATURE_CMO)); + + if (!firmware_has_feature(FW_FEATURE_CMO)) + return; + + for_each_possible_cpu(cpu) { + cmo_faults += be64_to_cpu(lppaca_of(cpu).cmo_faults); + cmo_fault_time += be64_to_cpu(lppaca_of(cpu).cmo_fault_time); + } + + seq_printf(m, "cmo_faults=%lu\n", cmo_faults); + seq_printf(m, "cmo_fault_time_usec=%lu\n", + cmo_fault_time / tb_ticks_per_usec); + seq_printf(m, "cmo_primary_psp=%d\n", cmo_get_primary_psp()); + seq_printf(m, "cmo_secondary_psp=%d\n", cmo_get_secondary_psp()); + seq_printf(m, "cmo_page_size=%lu\n", cmo_get_page_size()); +} + +static void splpar_dispatch_data(struct seq_file *m) +{ + int cpu; + unsigned long dispatches = 0; + unsigned long dispatch_dispersions = 0; + + for_each_possible_cpu(cpu) { + dispatches += be32_to_cpu(lppaca_of(cpu).yield_count); + dispatch_dispersions += + be32_to_cpu(lppaca_of(cpu).dispersion_count); + } + + seq_printf(m, "dispatches=%lu\n", dispatches); + seq_printf(m, "dispatch_dispersions=%lu\n", dispatch_dispersions); +} + +static void parse_em_data(struct seq_file *m) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE]; + + if (firmware_has_feature(FW_FEATURE_LPAR) && + plpar_hcall(H_GET_EM_PARMS, retbuf) == H_SUCCESS) + seq_printf(m, "power_mode_data=%016lx\n", retbuf[0]); +} + +static void maxmem_data(struct seq_file *m) +{ + unsigned long maxmem = 0; + + maxmem += (unsigned long)drmem_info->n_lmbs * drmem_info->lmb_size; + maxmem += hugetlb_total_pages() * PAGE_SIZE; + + seq_printf(m, "MaxMem=%lu\n", maxmem); +} + +static int pseries_lparcfg_data(struct seq_file *m, void *v) +{ + int partition_potential_processors; + int partition_active_processors; + struct device_node *rtas_node; + const __be32 *lrdrp = NULL; + + rtas_node = of_find_node_by_path("/rtas"); + if (rtas_node) + lrdrp = of_get_property(rtas_node, "ibm,lrdr-capacity", NULL); + + if (lrdrp == NULL) { + partition_potential_processors = vdso_data->processorCount; + } else { + partition_potential_processors = be32_to_cpup(lrdrp + 4); + } + of_node_put(rtas_node); + + partition_active_processors = lparcfg_count_active_processors(); + + if (firmware_has_feature(FW_FEATURE_SPLPAR)) { + /* this call handles the ibm,get-system-parameter contents */ + read_lpar_name(m); + parse_system_parameter_string(m); + parse_ppp_data(m); + parse_mpp_data(m); + parse_mpp_x_data(m); + pseries_cmo_data(m); + splpar_dispatch_data(m); + + seq_printf(m, "purr=%ld\n", get_purr()); + seq_printf(m, "tbr=%ld\n", mftb()); + } else { /* non SPLPAR case */ + + seq_printf(m, "system_active_processors=%d\n", + partition_potential_processors); + + seq_printf(m, "system_potential_processors=%d\n", + partition_potential_processors); + + seq_printf(m, "partition_max_entitled_capacity=%d\n", + partition_potential_processors * 100); + + seq_printf(m, "partition_entitled_capacity=%d\n", + partition_active_processors * 100); + } + + show_gpci_data(m); + + seq_printf(m, "partition_active_processors=%d\n", + partition_active_processors); + + seq_printf(m, "partition_potential_processors=%d\n", + partition_potential_processors); + + seq_printf(m, "shared_processor_mode=%d\n", + lppaca_shared_proc()); + +#ifdef CONFIG_PPC_64S_HASH_MMU + if (!radix_enabled()) + seq_printf(m, "slb_size=%d\n", mmu_slb_size); +#endif + parse_em_data(m); + maxmem_data(m); + + seq_printf(m, "security_flavor=%u\n", pseries_security_flavor); + + return 0; +} + +static ssize_t update_ppp(u64 *entitlement, u8 *weight) +{ + struct hvcall_ppp_data ppp_data; + u8 new_weight; + u64 new_entitled; + ssize_t retval; + + /* Get our current parameters */ + retval = h_get_ppp(&ppp_data); + if (retval) + return retval; + + if (entitlement) { + new_weight = ppp_data.weight; + new_entitled = *entitlement; + } else if (weight) { + new_weight = *weight; + new_entitled = ppp_data.entitlement; + } else + return -EINVAL; + + pr_debug("%s: current_entitled = %llu, current_weight = %u\n", + __func__, ppp_data.entitlement, ppp_data.weight); + + pr_debug("%s: new_entitled = %llu, new_weight = %u\n", + __func__, new_entitled, new_weight); + + retval = plpar_hcall_norets(H_SET_PPP, new_entitled, new_weight); + return retval; +} + +/** + * update_mpp + * + * Update the memory entitlement and weight for the partition. Caller must + * specify either a new entitlement or weight, not both, to be updated + * since the h_set_mpp call takes both entitlement and weight as parameters. + */ +static ssize_t update_mpp(u64 *entitlement, u8 *weight) +{ + struct hvcall_mpp_data mpp_data; + u64 new_entitled; + u8 new_weight; + ssize_t rc; + + if (entitlement) { + /* Check with vio to ensure the new memory entitlement + * can be handled. + */ + rc = vio_cmo_entitlement_update(*entitlement); + if (rc) + return rc; + } + + rc = h_get_mpp(&mpp_data); + if (rc) + return rc; + + if (entitlement) { + new_weight = mpp_data.mem_weight; + new_entitled = *entitlement; + } else if (weight) { + new_weight = *weight; + new_entitled = mpp_data.entitled_mem; + } else + return -EINVAL; + + pr_debug("%s: current_entitled = %lu, current_weight = %u\n", + __func__, mpp_data.entitled_mem, mpp_data.mem_weight); + + pr_debug("%s: new_entitled = %llu, new_weight = %u\n", + __func__, new_entitled, new_weight); + + rc = plpar_hcall_norets(H_SET_MPP, new_entitled, new_weight); + return rc; +} + +/* + * Interface for changing system parameters (variable capacity weight + * and entitled capacity). Format of input is "param_name=value"; + * anything after value is ignored. Valid parameters at this time are + * "partition_entitled_capacity" and "capacity_weight". We use + * H_SET_PPP to alter parameters. + * + * This function should be invoked only on systems with + * FW_FEATURE_SPLPAR. + */ +static ssize_t lparcfg_write(struct file *file, const char __user * buf, + size_t count, loff_t * off) +{ + char kbuf[64]; + char *tmp; + u64 new_entitled, *new_entitled_ptr = &new_entitled; + u8 new_weight, *new_weight_ptr = &new_weight; + ssize_t retval; + + if (!firmware_has_feature(FW_FEATURE_SPLPAR)) + return -EINVAL; + + if (count > sizeof(kbuf)) + return -EINVAL; + + if (copy_from_user(kbuf, buf, count)) + return -EFAULT; + + kbuf[count - 1] = '\0'; + tmp = strchr(kbuf, '='); + if (!tmp) + return -EINVAL; + + *tmp++ = '\0'; + + if (!strcmp(kbuf, "partition_entitled_capacity")) { + char *endp; + *new_entitled_ptr = (u64) simple_strtoul(tmp, &endp, 10); + if (endp == tmp) + return -EINVAL; + + retval = update_ppp(new_entitled_ptr, NULL); + + if (retval == H_SUCCESS || retval == H_CONSTRAINED) { + /* + * The hypervisor assigns VAS resources based + * on entitled capacity for shared mode. + * Reconfig VAS windows based on DLPAR CPU events. + */ + if (pseries_vas_dlpar_cpu() != 0) + retval = H_HARDWARE; + } + } else if (!strcmp(kbuf, "capacity_weight")) { + char *endp; + *new_weight_ptr = (u8) simple_strtoul(tmp, &endp, 10); + if (endp == tmp) + return -EINVAL; + + retval = update_ppp(NULL, new_weight_ptr); + } else if (!strcmp(kbuf, "entitled_memory")) { + char *endp; + *new_entitled_ptr = (u64) simple_strtoul(tmp, &endp, 10); + if (endp == tmp) + return -EINVAL; + + retval = update_mpp(new_entitled_ptr, NULL); + } else if (!strcmp(kbuf, "entitled_memory_weight")) { + char *endp; + *new_weight_ptr = (u8) simple_strtoul(tmp, &endp, 10); + if (endp == tmp) + return -EINVAL; + + retval = update_mpp(NULL, new_weight_ptr); + } else + return -EINVAL; + + if (retval == H_SUCCESS || retval == H_CONSTRAINED) { + retval = count; + } else if (retval == H_BUSY) { + retval = -EBUSY; + } else if (retval == H_HARDWARE) { + retval = -EIO; + } else if (retval == H_PARAMETER) { + retval = -EINVAL; + } + + return retval; +} + +static int lparcfg_data(struct seq_file *m, void *v) +{ + struct device_node *rootdn; + const char *model = ""; + const char *system_id = ""; + const char *tmp; + const __be32 *lp_index_ptr; + unsigned int lp_index = 0; + + seq_printf(m, "%s %s\n", MODULE_NAME, MODULE_VERS); + + rootdn = of_find_node_by_path("/"); + if (rootdn) { + tmp = of_get_property(rootdn, "model", NULL); + if (tmp) + model = tmp; + tmp = of_get_property(rootdn, "system-id", NULL); + if (tmp) + system_id = tmp; + lp_index_ptr = of_get_property(rootdn, "ibm,partition-no", + NULL); + if (lp_index_ptr) + lp_index = be32_to_cpup(lp_index_ptr); + of_node_put(rootdn); + } + seq_printf(m, "serial_number=%s\n", system_id); + seq_printf(m, "system_type=%s\n", model); + seq_printf(m, "partition_id=%d\n", (int)lp_index); + + return pseries_lparcfg_data(m, v); +} + +static int lparcfg_open(struct inode *inode, struct file *file) +{ + return single_open(file, lparcfg_data, NULL); +} + +static const struct proc_ops lparcfg_proc_ops = { + .proc_read = seq_read, + .proc_write = lparcfg_write, + .proc_open = lparcfg_open, + .proc_release = single_release, + .proc_lseek = seq_lseek, +}; + +static int __init lparcfg_init(void) +{ + umode_t mode = 0444; + + /* Allow writing if we have FW_FEATURE_SPLPAR */ + if (firmware_has_feature(FW_FEATURE_SPLPAR)) + mode |= 0200; + + if (!proc_create("powerpc/lparcfg", mode, NULL, &lparcfg_proc_ops)) { + printk(KERN_ERR "Failed to create powerpc/lparcfg\n"); + return -EIO; + } + return 0; +} +machine_device_initcall(pseries, lparcfg_init); diff --git a/arch/powerpc/platforms/pseries/mobility.c b/arch/powerpc/platforms/pseries/mobility.c new file mode 100644 index 000000000..0161226d8 --- /dev/null +++ b/arch/powerpc/platforms/pseries/mobility.c @@ -0,0 +1,830 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Support for Partition Mobility/Migration + * + * Copyright (C) 2010 Nathan Fontenot + * Copyright (C) 2010 IBM Corporation + */ + + +#define pr_fmt(fmt) "mobility: " fmt + +#include <linux/cpu.h> +#include <linux/kernel.h> +#include <linux/kobject.h> +#include <linux/nmi.h> +#include <linux/sched.h> +#include <linux/smp.h> +#include <linux/stat.h> +#include <linux/stop_machine.h> +#include <linux/completion.h> +#include <linux/device.h> +#include <linux/delay.h> +#include <linux/slab.h> +#include <linux/stringify.h> + +#include <asm/machdep.h> +#include <asm/nmi.h> +#include <asm/rtas.h> +#include "pseries.h" +#include "vas.h" /* vas_migration_handler() */ +#include "../../kernel/cacheinfo.h" + +static struct kobject *mobility_kobj; + +struct update_props_workarea { + __be32 phandle; + __be32 state; + __be64 reserved; + __be32 nprops; +} __packed; + +#define NODE_ACTION_MASK 0xff000000 +#define NODE_COUNT_MASK 0x00ffffff + +#define DELETE_DT_NODE 0x01000000 +#define UPDATE_DT_NODE 0x02000000 +#define ADD_DT_NODE 0x03000000 + +#define MIGRATION_SCOPE (1) +#define PRRN_SCOPE -2 + +#ifdef CONFIG_PPC_WATCHDOG +static unsigned int nmi_wd_lpm_factor = 200; + +#ifdef CONFIG_SYSCTL +static struct ctl_table nmi_wd_lpm_factor_ctl_table[] = { + { + .procname = "nmi_wd_lpm_factor", + .data = &nmi_wd_lpm_factor, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_douintvec_minmax, + }, + {} +}; + +static int __init register_nmi_wd_lpm_factor_sysctl(void) +{ + register_sysctl("kernel", nmi_wd_lpm_factor_ctl_table); + + return 0; +} +device_initcall(register_nmi_wd_lpm_factor_sysctl); +#endif /* CONFIG_SYSCTL */ +#endif /* CONFIG_PPC_WATCHDOG */ + +static int mobility_rtas_call(int token, char *buf, s32 scope) +{ + int rc; + + spin_lock(&rtas_data_buf_lock); + + memcpy(rtas_data_buf, buf, RTAS_DATA_BUF_SIZE); + rc = rtas_call(token, 2, 1, NULL, rtas_data_buf, scope); + memcpy(buf, rtas_data_buf, RTAS_DATA_BUF_SIZE); + + spin_unlock(&rtas_data_buf_lock); + return rc; +} + +static int delete_dt_node(struct device_node *dn) +{ + struct device_node *pdn; + bool is_platfac; + + pdn = of_get_parent(dn); + is_platfac = of_node_is_type(dn, "ibm,platform-facilities") || + of_node_is_type(pdn, "ibm,platform-facilities"); + of_node_put(pdn); + + /* + * The drivers that bind to nodes in the platform-facilities + * hierarchy don't support node removal, and the removal directive + * from firmware is always followed by an add of an equivalent + * node. The capability (e.g. RNG, encryption, compression) + * represented by the node is never interrupted by the migration. + * So ignore changes to this part of the tree. + */ + if (is_platfac) { + pr_notice("ignoring remove operation for %pOFfp\n", dn); + return 0; + } + + pr_debug("removing node %pOFfp\n", dn); + dlpar_detach_node(dn); + return 0; +} + +static int update_dt_property(struct device_node *dn, struct property **prop, + const char *name, u32 vd, char *value) +{ + struct property *new_prop = *prop; + int more = 0; + + /* A negative 'vd' value indicates that only part of the new property + * value is contained in the buffer and we need to call + * ibm,update-properties again to get the rest of the value. + * + * A negative value is also the two's compliment of the actual value. + */ + if (vd & 0x80000000) { + vd = ~vd + 1; + more = 1; + } + + if (new_prop) { + /* partial property fixup */ + char *new_data = kzalloc(new_prop->length + vd, GFP_KERNEL); + if (!new_data) + return -ENOMEM; + + memcpy(new_data, new_prop->value, new_prop->length); + memcpy(new_data + new_prop->length, value, vd); + + kfree(new_prop->value); + new_prop->value = new_data; + new_prop->length += vd; + } else { + new_prop = kzalloc(sizeof(*new_prop), GFP_KERNEL); + if (!new_prop) + return -ENOMEM; + + new_prop->name = kstrdup(name, GFP_KERNEL); + if (!new_prop->name) { + kfree(new_prop); + return -ENOMEM; + } + + new_prop->length = vd; + new_prop->value = kzalloc(new_prop->length, GFP_KERNEL); + if (!new_prop->value) { + kfree(new_prop->name); + kfree(new_prop); + return -ENOMEM; + } + + memcpy(new_prop->value, value, vd); + *prop = new_prop; + } + + if (!more) { + pr_debug("updating node %pOF property %s\n", dn, name); + of_update_property(dn, new_prop); + *prop = NULL; + } + + return 0; +} + +static int update_dt_node(struct device_node *dn, s32 scope) +{ + struct update_props_workarea *upwa; + struct property *prop = NULL; + int i, rc, rtas_rc; + char *prop_data; + char *rtas_buf; + int update_properties_token; + u32 nprops; + u32 vd; + + update_properties_token = rtas_function_token(RTAS_FN_IBM_UPDATE_PROPERTIES); + if (update_properties_token == RTAS_UNKNOWN_SERVICE) + return -EINVAL; + + rtas_buf = kzalloc(RTAS_DATA_BUF_SIZE, GFP_KERNEL); + if (!rtas_buf) + return -ENOMEM; + + upwa = (struct update_props_workarea *)&rtas_buf[0]; + upwa->phandle = cpu_to_be32(dn->phandle); + + do { + rtas_rc = mobility_rtas_call(update_properties_token, rtas_buf, + scope); + if (rtas_rc < 0) + break; + + prop_data = rtas_buf + sizeof(*upwa); + nprops = be32_to_cpu(upwa->nprops); + + /* On the first call to ibm,update-properties for a node the + * first property value descriptor contains an empty + * property name, the property value length encoded as u32, + * and the property value is the node path being updated. + */ + if (*prop_data == 0) { + prop_data++; + vd = be32_to_cpu(*(__be32 *)prop_data); + prop_data += vd + sizeof(vd); + nprops--; + } + + for (i = 0; i < nprops; i++) { + char *prop_name; + + prop_name = prop_data; + prop_data += strlen(prop_name) + 1; + vd = be32_to_cpu(*(__be32 *)prop_data); + prop_data += sizeof(vd); + + switch (vd) { + case 0x00000000: + /* name only property, nothing to do */ + break; + + case 0x80000000: + of_remove_property(dn, of_find_property(dn, + prop_name, NULL)); + prop = NULL; + break; + + default: + rc = update_dt_property(dn, &prop, prop_name, + vd, prop_data); + if (rc) { + pr_err("updating %s property failed: %d\n", + prop_name, rc); + } + + prop_data += vd; + break; + } + + cond_resched(); + } + + cond_resched(); + } while (rtas_rc == 1); + + kfree(rtas_buf); + return 0; +} + +static int add_dt_node(struct device_node *parent_dn, __be32 drc_index) +{ + struct device_node *dn; + int rc; + + dn = dlpar_configure_connector(drc_index, parent_dn); + if (!dn) + return -ENOENT; + + /* + * Since delete_dt_node() ignores this node type, this is the + * necessary counterpart. We also know that a platform-facilities + * node returned from dlpar_configure_connector() has children + * attached, and dlpar_attach_node() only adds the parent, leaking + * the children. So ignore these on the add side for now. + */ + if (of_node_is_type(dn, "ibm,platform-facilities")) { + pr_notice("ignoring add operation for %pOF\n", dn); + dlpar_free_cc_nodes(dn); + return 0; + } + + rc = dlpar_attach_node(dn, parent_dn); + if (rc) + dlpar_free_cc_nodes(dn); + + pr_debug("added node %pOFfp\n", dn); + + return rc; +} + +static int pseries_devicetree_update(s32 scope) +{ + char *rtas_buf; + __be32 *data; + int update_nodes_token; + int rc; + + update_nodes_token = rtas_function_token(RTAS_FN_IBM_UPDATE_NODES); + if (update_nodes_token == RTAS_UNKNOWN_SERVICE) + return 0; + + rtas_buf = kzalloc(RTAS_DATA_BUF_SIZE, GFP_KERNEL); + if (!rtas_buf) + return -ENOMEM; + + do { + rc = mobility_rtas_call(update_nodes_token, rtas_buf, scope); + if (rc && rc != 1) + break; + + data = (__be32 *)rtas_buf + 4; + while (be32_to_cpu(*data) & NODE_ACTION_MASK) { + int i; + u32 action = be32_to_cpu(*data) & NODE_ACTION_MASK; + u32 node_count = be32_to_cpu(*data) & NODE_COUNT_MASK; + + data++; + + for (i = 0; i < node_count; i++) { + struct device_node *np; + __be32 phandle = *data++; + __be32 drc_index; + + np = of_find_node_by_phandle(be32_to_cpu(phandle)); + if (!np) { + pr_warn("Failed lookup: phandle 0x%x for action 0x%x\n", + be32_to_cpu(phandle), action); + continue; + } + + switch (action) { + case DELETE_DT_NODE: + delete_dt_node(np); + break; + case UPDATE_DT_NODE: + update_dt_node(np, scope); + break; + case ADD_DT_NODE: + drc_index = *data++; + add_dt_node(np, drc_index); + break; + } + + of_node_put(np); + cond_resched(); + } + } + + cond_resched(); + } while (rc == 1); + + kfree(rtas_buf); + return rc; +} + +void post_mobility_fixup(void) +{ + int rc; + + rtas_activate_firmware(); + + /* + * We don't want CPUs to go online/offline while the device + * tree is being updated. + */ + cpus_read_lock(); + + /* + * It's common for the destination firmware to replace cache + * nodes. Release all of the cacheinfo hierarchy's references + * before updating the device tree. + */ + cacheinfo_teardown(); + + rc = pseries_devicetree_update(MIGRATION_SCOPE); + if (rc) + pr_err("device tree update failed: %d\n", rc); + + cacheinfo_rebuild(); + + cpus_read_unlock(); + + /* Possibly switch to a new L1 flush type */ + pseries_setup_security_mitigations(); + + /* Reinitialise system information for hv-24x7 */ + read_24x7_sys_info(); + + return; +} + +static int poll_vasi_state(u64 handle, unsigned long *res) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE]; + long hvrc; + int ret; + + hvrc = plpar_hcall(H_VASI_STATE, retbuf, handle); + switch (hvrc) { + case H_SUCCESS: + ret = 0; + *res = retbuf[0]; + break; + case H_PARAMETER: + ret = -EINVAL; + break; + case H_FUNCTION: + ret = -EOPNOTSUPP; + break; + case H_HARDWARE: + default: + pr_err("unexpected H_VASI_STATE result %ld\n", hvrc); + ret = -EIO; + break; + } + return ret; +} + +static int wait_for_vasi_session_suspending(u64 handle) +{ + unsigned long state; + int ret; + + /* + * Wait for transition from H_VASI_ENABLED to + * H_VASI_SUSPENDING. Treat anything else as an error. + */ + while (true) { + ret = poll_vasi_state(handle, &state); + + if (ret != 0 || state == H_VASI_SUSPENDING) { + break; + } else if (state == H_VASI_ENABLED) { + ssleep(1); + } else { + pr_err("unexpected H_VASI_STATE result %lu\n", state); + ret = -EIO; + break; + } + } + + /* + * Proceed even if H_VASI_STATE is unavailable. If H_JOIN or + * ibm,suspend-me are also unimplemented, we'll recover then. + */ + if (ret == -EOPNOTSUPP) + ret = 0; + + return ret; +} + +static void wait_for_vasi_session_completed(u64 handle) +{ + unsigned long state = 0; + int ret; + + pr_info("waiting for memory transfer to complete...\n"); + + /* + * Wait for transition from H_VASI_RESUMED to H_VASI_COMPLETED. + */ + while (true) { + ret = poll_vasi_state(handle, &state); + + /* + * If the memory transfer is already complete and the migration + * has been cleaned up by the hypervisor, H_PARAMETER is return, + * which is translate in EINVAL by poll_vasi_state(). + */ + if (ret == -EINVAL || (!ret && state == H_VASI_COMPLETED)) { + pr_info("memory transfer completed.\n"); + break; + } + + if (ret) { + pr_err("H_VASI_STATE return error (%d)\n", ret); + break; + } + + if (state != H_VASI_RESUMED) { + pr_err("unexpected H_VASI_STATE result %lu\n", state); + break; + } + + msleep(500); + } +} + +static void prod_single(unsigned int target_cpu) +{ + long hvrc; + int hwid; + + hwid = get_hard_smp_processor_id(target_cpu); + hvrc = plpar_hcall_norets(H_PROD, hwid); + if (hvrc == H_SUCCESS) + return; + pr_err_ratelimited("H_PROD of CPU %u (hwid %d) error: %ld\n", + target_cpu, hwid, hvrc); +} + +static void prod_others(void) +{ + unsigned int cpu; + + for_each_online_cpu(cpu) { + if (cpu != smp_processor_id()) + prod_single(cpu); + } +} + +static u16 clamp_slb_size(void) +{ +#ifdef CONFIG_PPC_64S_HASH_MMU + u16 prev = mmu_slb_size; + + slb_set_size(SLB_MIN_SIZE); + + return prev; +#else + return 0; +#endif +} + +static int do_suspend(void) +{ + u16 saved_slb_size; + int status; + int ret; + + pr_info("calling ibm,suspend-me on CPU %i\n", smp_processor_id()); + + /* + * The destination processor model may have fewer SLB entries + * than the source. We reduce mmu_slb_size to a safe minimum + * before suspending in order to minimize the possibility of + * programming non-existent entries on the destination. If + * suspend fails, we restore it before returning. On success + * the OF reconfig path will update it from the new device + * tree after resuming on the destination. + */ + saved_slb_size = clamp_slb_size(); + + ret = rtas_ibm_suspend_me(&status); + if (ret != 0) { + pr_err("ibm,suspend-me error: %d\n", status); + slb_set_size(saved_slb_size); + } + + return ret; +} + +/** + * struct pseries_suspend_info - State shared between CPUs for join/suspend. + * @counter: Threads are to increment this upon resuming from suspend + * or if an error is received from H_JOIN. The thread which performs + * the first increment (i.e. sets it to 1) is responsible for + * waking the other threads. + * @done: False if join/suspend is in progress. True if the operation is + * complete (successful or not). + */ +struct pseries_suspend_info { + atomic_t counter; + bool done; +}; + +static int do_join(void *arg) +{ + struct pseries_suspend_info *info = arg; + atomic_t *counter = &info->counter; + long hvrc; + int ret; + +retry: + /* Must ensure MSR.EE off for H_JOIN. */ + hard_irq_disable(); + hvrc = plpar_hcall_norets(H_JOIN); + + switch (hvrc) { + case H_CONTINUE: + /* + * All other CPUs are offline or in H_JOIN. This CPU + * attempts the suspend. + */ + ret = do_suspend(); + break; + case H_SUCCESS: + /* + * The suspend is complete and this cpu has received a + * prod, or we've received a stray prod from unrelated + * code (e.g. paravirt spinlocks) and we need to join + * again. + * + * This barrier orders the return from H_JOIN above vs + * the load of info->done. It pairs with the barrier + * in the wakeup/prod path below. + */ + smp_mb(); + if (READ_ONCE(info->done) == false) { + pr_info_ratelimited("premature return from H_JOIN on CPU %i, retrying", + smp_processor_id()); + goto retry; + } + ret = 0; + break; + case H_BAD_MODE: + case H_HARDWARE: + default: + ret = -EIO; + pr_err_ratelimited("H_JOIN error %ld on CPU %i\n", + hvrc, smp_processor_id()); + break; + } + + if (atomic_inc_return(counter) == 1) { + pr_info("CPU %u waking all threads\n", smp_processor_id()); + WRITE_ONCE(info->done, true); + /* + * This barrier orders the store to info->done vs subsequent + * H_PRODs to wake the other CPUs. It pairs with the barrier + * in the H_SUCCESS case above. + */ + smp_mb(); + prod_others(); + } + /* + * Execution may have been suspended for several seconds, so reset + * the watchdogs. touch_nmi_watchdog() also touches the soft lockup + * watchdog. + */ + rcu_cpu_stall_reset(); + touch_nmi_watchdog(); + + return ret; +} + +/* + * Abort reason code byte 0. We use only the 'Migrating partition' value. + */ +enum vasi_aborting_entity { + ORCHESTRATOR = 1, + VSP_SOURCE = 2, + PARTITION_FIRMWARE = 3, + PLATFORM_FIRMWARE = 4, + VSP_TARGET = 5, + MIGRATING_PARTITION = 6, +}; + +static void pseries_cancel_migration(u64 handle, int err) +{ + u32 reason_code; + u32 detail; + u8 entity; + long hvrc; + + entity = MIGRATING_PARTITION; + detail = abs(err) & 0xffffff; + reason_code = (entity << 24) | detail; + + hvrc = plpar_hcall_norets(H_VASI_SIGNAL, handle, + H_VASI_SIGNAL_CANCEL, reason_code); + if (hvrc) + pr_err("H_VASI_SIGNAL error: %ld\n", hvrc); +} + +static int pseries_suspend(u64 handle) +{ + const unsigned int max_attempts = 5; + unsigned int retry_interval_ms = 1; + unsigned int attempt = 1; + int ret; + + while (true) { + struct pseries_suspend_info info; + unsigned long vasi_state; + int vasi_err; + + info = (struct pseries_suspend_info) { + .counter = ATOMIC_INIT(0), + .done = false, + }; + + ret = stop_machine(do_join, &info, cpu_online_mask); + if (ret == 0) + break; + /* + * Encountered an error. If the VASI stream is still + * in Suspending state, it's likely a transient + * condition related to some device in the partition + * and we can retry in the hope that the cause has + * cleared after some delay. + * + * A better design would allow drivers etc to prepare + * for the suspend and avoid conditions which prevent + * the suspend from succeeding. For now, we have this + * mitigation. + */ + pr_notice("Partition suspend attempt %u of %u error: %d\n", + attempt, max_attempts, ret); + + if (attempt == max_attempts) + break; + + vasi_err = poll_vasi_state(handle, &vasi_state); + if (vasi_err == 0) { + if (vasi_state != H_VASI_SUSPENDING) { + pr_notice("VASI state %lu after failed suspend\n", + vasi_state); + break; + } + } else if (vasi_err != -EOPNOTSUPP) { + pr_err("VASI state poll error: %d", vasi_err); + break; + } + + pr_notice("Will retry partition suspend after %u ms\n", + retry_interval_ms); + + msleep(retry_interval_ms); + retry_interval_ms *= 10; + attempt++; + } + + return ret; +} + +static int pseries_migrate_partition(u64 handle) +{ + int ret; + unsigned int factor = 0; + +#ifdef CONFIG_PPC_WATCHDOG + factor = nmi_wd_lpm_factor; +#endif + /* + * When the migration is initiated, the hypervisor changes VAS + * mappings to prepare before OS gets the notification and + * closes all VAS windows. NX generates continuous faults during + * this time and the user space can not differentiate these + * faults from the migration event. So reduce this time window + * by closing VAS windows at the beginning of this function. + */ + vas_migration_handler(VAS_SUSPEND); + + ret = wait_for_vasi_session_suspending(handle); + if (ret) + goto out; + + if (factor) + watchdog_hardlockup_set_timeout_pct(factor); + + ret = pseries_suspend(handle); + if (ret == 0) { + post_mobility_fixup(); + /* + * Wait until the memory transfer is complete, so that the user + * space process returns from the syscall after the transfer is + * complete. This allows the user hooks to be executed at the + * right time. + */ + wait_for_vasi_session_completed(handle); + } else + pseries_cancel_migration(handle, ret); + + if (factor) + watchdog_hardlockup_set_timeout_pct(0); + +out: + vas_migration_handler(VAS_RESUME); + + return ret; +} + +int rtas_syscall_dispatch_ibm_suspend_me(u64 handle) +{ + return pseries_migrate_partition(handle); +} + +static ssize_t migration_store(const struct class *class, + const struct class_attribute *attr, const char *buf, + size_t count) +{ + u64 streamid; + int rc; + + rc = kstrtou64(buf, 0, &streamid); + if (rc) + return rc; + + rc = pseries_migrate_partition(streamid); + if (rc) + return rc; + + return count; +} + +/* + * Used by drmgr to determine the kernel behavior of the migration interface. + * + * Version 1: Performs all PAPR requirements for migration including + * firmware activation and device tree update. + */ +#define MIGRATION_API_VERSION 1 + +static CLASS_ATTR_WO(migration); +static CLASS_ATTR_STRING(api_version, 0444, __stringify(MIGRATION_API_VERSION)); + +static int __init mobility_sysfs_init(void) +{ + int rc; + + mobility_kobj = kobject_create_and_add("mobility", kernel_kobj); + if (!mobility_kobj) + return -ENOMEM; + + rc = sysfs_create_file(mobility_kobj, &class_attr_migration.attr); + if (rc) + pr_err("unable to create migration sysfs file (%d)\n", rc); + + rc = sysfs_create_file(mobility_kobj, &class_attr_api_version.attr.attr); + if (rc) + pr_err("unable to create api_version sysfs file (%d)\n", rc); + + return 0; +} +machine_device_initcall(pseries, mobility_sysfs_init); diff --git a/arch/powerpc/platforms/pseries/msi.c b/arch/powerpc/platforms/pseries/msi.c new file mode 100644 index 000000000..423ee1d5b --- /dev/null +++ b/arch/powerpc/platforms/pseries/msi.c @@ -0,0 +1,698 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright 2006 Jake Moilanen <moilanen@austin.ibm.com>, IBM Corp. + * Copyright 2006-2007 Michael Ellerman, IBM Corp. + */ + +#include <linux/crash_dump.h> +#include <linux/device.h> +#include <linux/irq.h> +#include <linux/irqdomain.h> +#include <linux/msi.h> + +#include <asm/rtas.h> +#include <asm/hw_irq.h> +#include <asm/ppc-pci.h> +#include <asm/machdep.h> +#include <asm/xive.h> + +#include "pseries.h" + +static int query_token, change_token; + +#define RTAS_QUERY_FN 0 +#define RTAS_CHANGE_FN 1 +#define RTAS_RESET_FN 2 +#define RTAS_CHANGE_MSI_FN 3 +#define RTAS_CHANGE_MSIX_FN 4 +#define RTAS_CHANGE_32MSI_FN 5 + +/* RTAS Helpers */ + +static int rtas_change_msi(struct pci_dn *pdn, u32 func, u32 num_irqs) +{ + u32 addr, seq_num, rtas_ret[3]; + unsigned long buid; + int rc; + + addr = rtas_config_addr(pdn->busno, pdn->devfn, 0); + buid = pdn->phb->buid; + + seq_num = 1; + do { + if (func == RTAS_CHANGE_MSI_FN || func == RTAS_CHANGE_MSIX_FN || + func == RTAS_CHANGE_32MSI_FN) + rc = rtas_call(change_token, 6, 4, rtas_ret, addr, + BUID_HI(buid), BUID_LO(buid), + func, num_irqs, seq_num); + else + rc = rtas_call(change_token, 6, 3, rtas_ret, addr, + BUID_HI(buid), BUID_LO(buid), + func, num_irqs, seq_num); + + seq_num = rtas_ret[1]; + } while (rtas_busy_delay(rc)); + + /* + * If the RTAS call succeeded, return the number of irqs allocated. + * If not, make sure we return a negative error code. + */ + if (rc == 0) + rc = rtas_ret[0]; + else if (rc > 0) + rc = -rc; + + pr_debug("rtas_msi: ibm,change_msi(func=%d,num=%d), got %d rc = %d\n", + func, num_irqs, rtas_ret[0], rc); + + return rc; +} + +static void rtas_disable_msi(struct pci_dev *pdev) +{ + struct pci_dn *pdn; + + pdn = pci_get_pdn(pdev); + if (!pdn) + return; + + /* + * disabling MSI with the explicit interface also disables MSI-X + */ + if (rtas_change_msi(pdn, RTAS_CHANGE_MSI_FN, 0) != 0) { + /* + * may have failed because explicit interface is not + * present + */ + if (rtas_change_msi(pdn, RTAS_CHANGE_FN, 0) != 0) { + pr_debug("rtas_msi: Setting MSIs to 0 failed!\n"); + } + } +} + +static int rtas_query_irq_number(struct pci_dn *pdn, int offset) +{ + u32 addr, rtas_ret[2]; + unsigned long buid; + int rc; + + addr = rtas_config_addr(pdn->busno, pdn->devfn, 0); + buid = pdn->phb->buid; + + do { + rc = rtas_call(query_token, 4, 3, rtas_ret, addr, + BUID_HI(buid), BUID_LO(buid), offset); + } while (rtas_busy_delay(rc)); + + if (rc) { + pr_debug("rtas_msi: error (%d) querying source number\n", rc); + return rc; + } + + return rtas_ret[0]; +} + +static int check_req(struct pci_dev *pdev, int nvec, char *prop_name) +{ + struct device_node *dn; + const __be32 *p; + u32 req_msi; + + dn = pci_device_to_OF_node(pdev); + + p = of_get_property(dn, prop_name, NULL); + if (!p) { + pr_debug("rtas_msi: No %s on %pOF\n", prop_name, dn); + return -ENOENT; + } + + req_msi = be32_to_cpup(p); + if (req_msi < nvec) { + pr_debug("rtas_msi: %s requests < %d MSIs\n", prop_name, nvec); + + if (req_msi == 0) /* Be paranoid */ + return -ENOSPC; + + return req_msi; + } + + return 0; +} + +static int check_req_msi(struct pci_dev *pdev, int nvec) +{ + return check_req(pdev, nvec, "ibm,req#msi"); +} + +static int check_req_msix(struct pci_dev *pdev, int nvec) +{ + return check_req(pdev, nvec, "ibm,req#msi-x"); +} + +/* Quota calculation */ + +static struct device_node *__find_pe_total_msi(struct device_node *node, int *total) +{ + struct device_node *dn; + const __be32 *p; + + dn = of_node_get(node); + while (dn) { + p = of_get_property(dn, "ibm,pe-total-#msi", NULL); + if (p) { + pr_debug("rtas_msi: found prop on dn %pOF\n", + dn); + *total = be32_to_cpup(p); + return dn; + } + + dn = of_get_next_parent(dn); + } + + return NULL; +} + +static struct device_node *find_pe_total_msi(struct pci_dev *dev, int *total) +{ + return __find_pe_total_msi(pci_device_to_OF_node(dev), total); +} + +static struct device_node *find_pe_dn(struct pci_dev *dev, int *total) +{ + struct device_node *dn; + struct eeh_dev *edev; + + /* Found our PE and assume 8 at that point. */ + + dn = pci_device_to_OF_node(dev); + if (!dn) + return NULL; + + /* Get the top level device in the PE */ + edev = pdn_to_eeh_dev(PCI_DN(dn)); + if (edev->pe) + edev = list_first_entry(&edev->pe->edevs, struct eeh_dev, + entry); + dn = pci_device_to_OF_node(edev->pdev); + if (!dn) + return NULL; + + /* We actually want the parent */ + dn = of_get_parent(dn); + if (!dn) + return NULL; + + /* Hardcode of 8 for old firmwares */ + *total = 8; + pr_debug("rtas_msi: using PE dn %pOF\n", dn); + + return dn; +} + +struct msi_counts { + struct device_node *requestor; + int num_devices; + int request; + int quota; + int spare; + int over_quota; +}; + +static void *count_non_bridge_devices(struct device_node *dn, void *data) +{ + struct msi_counts *counts = data; + const __be32 *p; + u32 class; + + pr_debug("rtas_msi: counting %pOF\n", dn); + + p = of_get_property(dn, "class-code", NULL); + class = p ? be32_to_cpup(p) : 0; + + if ((class >> 8) != PCI_CLASS_BRIDGE_PCI) + counts->num_devices++; + + return NULL; +} + +static void *count_spare_msis(struct device_node *dn, void *data) +{ + struct msi_counts *counts = data; + const __be32 *p; + int req; + + if (dn == counts->requestor) + req = counts->request; + else { + /* We don't know if a driver will try to use MSI or MSI-X, + * so we just have to punt and use the larger of the two. */ + req = 0; + p = of_get_property(dn, "ibm,req#msi", NULL); + if (p) + req = be32_to_cpup(p); + + p = of_get_property(dn, "ibm,req#msi-x", NULL); + if (p) + req = max(req, (int)be32_to_cpup(p)); + } + + if (req < counts->quota) + counts->spare += counts->quota - req; + else if (req > counts->quota) + counts->over_quota++; + + return NULL; +} + +static int msi_quota_for_device(struct pci_dev *dev, int request) +{ + struct device_node *pe_dn; + struct msi_counts counts; + int total; + + pr_debug("rtas_msi: calc quota for %s, request %d\n", pci_name(dev), + request); + + pe_dn = find_pe_total_msi(dev, &total); + if (!pe_dn) + pe_dn = find_pe_dn(dev, &total); + + if (!pe_dn) { + pr_err("rtas_msi: couldn't find PE for %s\n", pci_name(dev)); + goto out; + } + + pr_debug("rtas_msi: found PE %pOF\n", pe_dn); + + memset(&counts, 0, sizeof(struct msi_counts)); + + /* Work out how many devices we have below this PE */ + pci_traverse_device_nodes(pe_dn, count_non_bridge_devices, &counts); + + if (counts.num_devices == 0) { + pr_err("rtas_msi: found 0 devices under PE for %s\n", + pci_name(dev)); + goto out; + } + + counts.quota = total / counts.num_devices; + if (request <= counts.quota) + goto out; + + /* else, we have some more calculating to do */ + counts.requestor = pci_device_to_OF_node(dev); + counts.request = request; + pci_traverse_device_nodes(pe_dn, count_spare_msis, &counts); + + /* If the quota isn't an integer multiple of the total, we can + * use the remainder as spare MSIs for anyone that wants them. */ + counts.spare += total % counts.num_devices; + + /* Divide any spare by the number of over-quota requestors */ + if (counts.over_quota) + counts.quota += counts.spare / counts.over_quota; + + /* And finally clamp the request to the possibly adjusted quota */ + request = min(counts.quota, request); + + pr_debug("rtas_msi: request clamped to quota %d\n", request); +out: + of_node_put(pe_dn); + + return request; +} + +static void rtas_hack_32bit_msi_gen2(struct pci_dev *pdev) +{ + u32 addr_hi, addr_lo; + + /* + * We should only get in here for IODA1 configs. This is based on the + * fact that we using RTAS for MSIs, we don't have the 32 bit MSI RTAS + * support, and we are in a PCIe Gen2 slot. + */ + dev_info(&pdev->dev, + "rtas_msi: No 32 bit MSI firmware support, forcing 32 bit MSI\n"); + pci_read_config_dword(pdev, pdev->msi_cap + PCI_MSI_ADDRESS_HI, &addr_hi); + addr_lo = 0xffff0000 | ((addr_hi >> (48 - 32)) << 4); + pci_write_config_dword(pdev, pdev->msi_cap + PCI_MSI_ADDRESS_LO, addr_lo); + pci_write_config_dword(pdev, pdev->msi_cap + PCI_MSI_ADDRESS_HI, 0); +} + +static int rtas_prepare_msi_irqs(struct pci_dev *pdev, int nvec_in, int type, + msi_alloc_info_t *arg) +{ + struct pci_dn *pdn; + int quota, rc; + int nvec = nvec_in; + int use_32bit_msi_hack = 0; + + if (type == PCI_CAP_ID_MSIX) + rc = check_req_msix(pdev, nvec); + else + rc = check_req_msi(pdev, nvec); + + if (rc) + return rc; + + quota = msi_quota_for_device(pdev, nvec); + + if (quota && quota < nvec) + return quota; + + /* + * Firmware currently refuse any non power of two allocation + * so we round up if the quota will allow it. + */ + if (type == PCI_CAP_ID_MSIX) { + int m = roundup_pow_of_two(nvec); + quota = msi_quota_for_device(pdev, m); + + if (quota >= m) + nvec = m; + } + + pdn = pci_get_pdn(pdev); + + /* + * Try the new more explicit firmware interface, if that fails fall + * back to the old interface. The old interface is known to never + * return MSI-Xs. + */ +again: + if (type == PCI_CAP_ID_MSI) { + if (pdev->no_64bit_msi) { + rc = rtas_change_msi(pdn, RTAS_CHANGE_32MSI_FN, nvec); + if (rc < 0) { + /* + * We only want to run the 32 bit MSI hack below if + * the max bus speed is Gen2 speed + */ + if (pdev->bus->max_bus_speed != PCIE_SPEED_5_0GT) + return rc; + + use_32bit_msi_hack = 1; + } + } else + rc = -1; + + if (rc < 0) + rc = rtas_change_msi(pdn, RTAS_CHANGE_MSI_FN, nvec); + + if (rc < 0) { + pr_debug("rtas_msi: trying the old firmware call.\n"); + rc = rtas_change_msi(pdn, RTAS_CHANGE_FN, nvec); + } + + if (use_32bit_msi_hack && rc > 0) + rtas_hack_32bit_msi_gen2(pdev); + } else + rc = rtas_change_msi(pdn, RTAS_CHANGE_MSIX_FN, nvec); + + if (rc != nvec) { + if (nvec != nvec_in) { + nvec = nvec_in; + goto again; + } + pr_debug("rtas_msi: rtas_change_msi() failed\n"); + return rc; + } + + return 0; +} + +static int pseries_msi_ops_prepare(struct irq_domain *domain, struct device *dev, + int nvec, msi_alloc_info_t *arg) +{ + struct pci_dev *pdev = to_pci_dev(dev); + int type = pdev->msix_enabled ? PCI_CAP_ID_MSIX : PCI_CAP_ID_MSI; + + return rtas_prepare_msi_irqs(pdev, nvec, type, arg); +} + +/* + * ->msi_free() is called before irq_domain_free_irqs_top() when the + * handler data is still available. Use that to clear the XIVE + * controller data. + */ +static void pseries_msi_ops_msi_free(struct irq_domain *domain, + struct msi_domain_info *info, + unsigned int irq) +{ + if (xive_enabled()) + xive_irq_free_data(irq); +} + +/* + * RTAS can not disable one MSI at a time. It's all or nothing. Do it + * at the end after all IRQs have been freed. + */ +static void pseries_msi_post_free(struct irq_domain *domain, struct device *dev) +{ + if (WARN_ON_ONCE(!dev_is_pci(dev))) + return; + + rtas_disable_msi(to_pci_dev(dev)); +} + +static struct msi_domain_ops pseries_pci_msi_domain_ops = { + .msi_prepare = pseries_msi_ops_prepare, + .msi_free = pseries_msi_ops_msi_free, + .msi_post_free = pseries_msi_post_free, +}; + +static void pseries_msi_shutdown(struct irq_data *d) +{ + d = d->parent_data; + if (d->chip->irq_shutdown) + d->chip->irq_shutdown(d); +} + +static void pseries_msi_mask(struct irq_data *d) +{ + pci_msi_mask_irq(d); + irq_chip_mask_parent(d); +} + +static void pseries_msi_unmask(struct irq_data *d) +{ + pci_msi_unmask_irq(d); + irq_chip_unmask_parent(d); +} + +static void pseries_msi_write_msg(struct irq_data *data, struct msi_msg *msg) +{ + struct msi_desc *entry = irq_data_get_msi_desc(data); + + /* + * Do not update the MSIx vector table. It's not strictly necessary + * because the table is initialized by the underlying hypervisor, PowerVM + * or QEMU/KVM. However, if the MSIx vector entry is cleared, any further + * activation will fail. This can happen in some drivers (eg. IPR) which + * deactivate an IRQ used for testing MSI support. + */ + entry->msg = *msg; +} + +static struct irq_chip pseries_pci_msi_irq_chip = { + .name = "pSeries-PCI-MSI", + .irq_shutdown = pseries_msi_shutdown, + .irq_mask = pseries_msi_mask, + .irq_unmask = pseries_msi_unmask, + .irq_eoi = irq_chip_eoi_parent, + .irq_write_msi_msg = pseries_msi_write_msg, +}; + + +/* + * Set MSI_FLAG_MSIX_CONTIGUOUS as there is no way to express to + * firmware to request a discontiguous or non-zero based range of + * MSI-X entries. Core code will reject such setup attempts. + */ +static struct msi_domain_info pseries_msi_domain_info = { + .flags = (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS | + MSI_FLAG_MULTI_PCI_MSI | MSI_FLAG_PCI_MSIX | + MSI_FLAG_MSIX_CONTIGUOUS), + .ops = &pseries_pci_msi_domain_ops, + .chip = &pseries_pci_msi_irq_chip, +}; + +static void pseries_msi_compose_msg(struct irq_data *data, struct msi_msg *msg) +{ + __pci_read_msi_msg(irq_data_get_msi_desc(data), msg); +} + +static struct irq_chip pseries_msi_irq_chip = { + .name = "pSeries-MSI", + .irq_shutdown = pseries_msi_shutdown, + .irq_mask = irq_chip_mask_parent, + .irq_unmask = irq_chip_unmask_parent, + .irq_eoi = irq_chip_eoi_parent, + .irq_set_affinity = irq_chip_set_affinity_parent, + .irq_compose_msi_msg = pseries_msi_compose_msg, +}; + +static int pseries_irq_parent_domain_alloc(struct irq_domain *domain, unsigned int virq, + irq_hw_number_t hwirq) +{ + struct irq_fwspec parent_fwspec; + int ret; + + parent_fwspec.fwnode = domain->parent->fwnode; + parent_fwspec.param_count = 2; + parent_fwspec.param[0] = hwirq; + parent_fwspec.param[1] = IRQ_TYPE_EDGE_RISING; + + ret = irq_domain_alloc_irqs_parent(domain, virq, 1, &parent_fwspec); + if (ret) + return ret; + + return 0; +} + +static int pseries_irq_domain_alloc(struct irq_domain *domain, unsigned int virq, + unsigned int nr_irqs, void *arg) +{ + struct pci_controller *phb = domain->host_data; + msi_alloc_info_t *info = arg; + struct msi_desc *desc = info->desc; + struct pci_dev *pdev = msi_desc_to_pci_dev(desc); + int hwirq; + int i, ret; + + hwirq = rtas_query_irq_number(pci_get_pdn(pdev), desc->msi_index); + if (hwirq < 0) { + dev_err(&pdev->dev, "Failed to query HW IRQ: %d\n", hwirq); + return hwirq; + } + + dev_dbg(&pdev->dev, "%s bridge %pOF %d/%x #%d\n", __func__, + phb->dn, virq, hwirq, nr_irqs); + + for (i = 0; i < nr_irqs; i++) { + ret = pseries_irq_parent_domain_alloc(domain, virq + i, hwirq + i); + if (ret) + goto out; + + irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i, + &pseries_msi_irq_chip, domain->host_data); + } + + return 0; + +out: + /* TODO: handle RTAS cleanup in ->msi_finish() ? */ + irq_domain_free_irqs_parent(domain, virq, i - 1); + return ret; +} + +static void pseries_irq_domain_free(struct irq_domain *domain, unsigned int virq, + unsigned int nr_irqs) +{ + struct irq_data *d = irq_domain_get_irq_data(domain, virq); + struct pci_controller *phb = irq_data_get_irq_chip_data(d); + + pr_debug("%s bridge %pOF %d #%d\n", __func__, phb->dn, virq, nr_irqs); + + /* XIVE domain data is cleared through ->msi_free() */ +} + +static const struct irq_domain_ops pseries_irq_domain_ops = { + .alloc = pseries_irq_domain_alloc, + .free = pseries_irq_domain_free, +}; + +static int __pseries_msi_allocate_domains(struct pci_controller *phb, + unsigned int count) +{ + struct irq_domain *parent = irq_get_default_host(); + + phb->fwnode = irq_domain_alloc_named_id_fwnode("pSeries-MSI", + phb->global_number); + if (!phb->fwnode) + return -ENOMEM; + + phb->dev_domain = irq_domain_create_hierarchy(parent, 0, count, + phb->fwnode, + &pseries_irq_domain_ops, phb); + if (!phb->dev_domain) { + pr_err("PCI: failed to create IRQ domain bridge %pOF (domain %d)\n", + phb->dn, phb->global_number); + irq_domain_free_fwnode(phb->fwnode); + return -ENOMEM; + } + + phb->msi_domain = pci_msi_create_irq_domain(of_node_to_fwnode(phb->dn), + &pseries_msi_domain_info, + phb->dev_domain); + if (!phb->msi_domain) { + pr_err("PCI: failed to create MSI IRQ domain bridge %pOF (domain %d)\n", + phb->dn, phb->global_number); + irq_domain_free_fwnode(phb->fwnode); + irq_domain_remove(phb->dev_domain); + return -ENOMEM; + } + + return 0; +} + +int pseries_msi_allocate_domains(struct pci_controller *phb) +{ + int count; + + if (!__find_pe_total_msi(phb->dn, &count)) { + pr_err("PCI: failed to find MSIs for bridge %pOF (domain %d)\n", + phb->dn, phb->global_number); + return -ENOSPC; + } + + return __pseries_msi_allocate_domains(phb, count); +} + +void pseries_msi_free_domains(struct pci_controller *phb) +{ + if (phb->msi_domain) + irq_domain_remove(phb->msi_domain); + if (phb->dev_domain) + irq_domain_remove(phb->dev_domain); + if (phb->fwnode) + irq_domain_free_fwnode(phb->fwnode); +} + +static void rtas_msi_pci_irq_fixup(struct pci_dev *pdev) +{ + /* No LSI -> leave MSIs (if any) configured */ + if (!pdev->irq) { + dev_dbg(&pdev->dev, "rtas_msi: no LSI, nothing to do.\n"); + return; + } + + /* No MSI -> MSIs can't have been assigned by fw, leave LSI */ + if (check_req_msi(pdev, 1) && check_req_msix(pdev, 1)) { + dev_dbg(&pdev->dev, "rtas_msi: no req#msi/x, nothing to do.\n"); + return; + } + + dev_dbg(&pdev->dev, "rtas_msi: disabling existing MSI.\n"); + rtas_disable_msi(pdev); +} + +static int rtas_msi_init(void) +{ + query_token = rtas_function_token(RTAS_FN_IBM_QUERY_INTERRUPT_SOURCE_NUMBER); + change_token = rtas_function_token(RTAS_FN_IBM_CHANGE_MSI); + + if ((query_token == RTAS_UNKNOWN_SERVICE) || + (change_token == RTAS_UNKNOWN_SERVICE)) { + pr_debug("rtas_msi: no RTAS tokens, no MSI support.\n"); + return -1; + } + + pr_debug("rtas_msi: Registering RTAS MSI callbacks.\n"); + + WARN_ON(ppc_md.pci_irq_fixup); + ppc_md.pci_irq_fixup = rtas_msi_pci_irq_fixup; + + return 0; +} +machine_arch_initcall(pseries, rtas_msi_init); diff --git a/arch/powerpc/platforms/pseries/nvram.c b/arch/powerpc/platforms/pseries/nvram.c new file mode 100644 index 000000000..8130c3796 --- /dev/null +++ b/arch/powerpc/platforms/pseries/nvram.c @@ -0,0 +1,241 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * c 2001 PPC 64 Team, IBM Corp + * + * /dev/nvram driver for PPC64 + */ + + +#include <linux/types.h> +#include <linux/errno.h> +#include <linux/init.h> +#include <linux/spinlock.h> +#include <linux/slab.h> +#include <linux/ctype.h> +#include <linux/uaccess.h> +#include <linux/of.h> +#include <asm/nvram.h> +#include <asm/rtas.h> +#include <asm/machdep.h> + +/* Max bytes to read/write in one go */ +#define NVRW_CNT 0x20 + +static unsigned int nvram_size; +static int nvram_fetch, nvram_store; +static char nvram_buf[NVRW_CNT]; /* assume this is in the first 4GB */ +static DEFINE_SPINLOCK(nvram_lock); + +/* See clobbering_unread_rtas_event() */ +#define NVRAM_RTAS_READ_TIMEOUT 5 /* seconds */ +static time64_t last_unread_rtas_event; /* timestamp */ + +#ifdef CONFIG_PSTORE +time64_t last_rtas_event; +#endif + +static ssize_t pSeries_nvram_read(char *buf, size_t count, loff_t *index) +{ + unsigned int i; + unsigned long len; + int done; + unsigned long flags; + char *p = buf; + + + if (nvram_size == 0 || nvram_fetch == RTAS_UNKNOWN_SERVICE) + return -ENODEV; + + if (*index >= nvram_size) + return 0; + + i = *index; + if (i + count > nvram_size) + count = nvram_size - i; + + spin_lock_irqsave(&nvram_lock, flags); + + for (; count != 0; count -= len) { + len = count; + if (len > NVRW_CNT) + len = NVRW_CNT; + + if ((rtas_call(nvram_fetch, 3, 2, &done, i, __pa(nvram_buf), + len) != 0) || len != done) { + spin_unlock_irqrestore(&nvram_lock, flags); + return -EIO; + } + + memcpy(p, nvram_buf, len); + + p += len; + i += len; + } + + spin_unlock_irqrestore(&nvram_lock, flags); + + *index = i; + return p - buf; +} + +static ssize_t pSeries_nvram_write(char *buf, size_t count, loff_t *index) +{ + unsigned int i; + unsigned long len; + int done; + unsigned long flags; + const char *p = buf; + + if (nvram_size == 0 || nvram_store == RTAS_UNKNOWN_SERVICE) + return -ENODEV; + + if (*index >= nvram_size) + return 0; + + i = *index; + if (i + count > nvram_size) + count = nvram_size - i; + + spin_lock_irqsave(&nvram_lock, flags); + + for (; count != 0; count -= len) { + len = count; + if (len > NVRW_CNT) + len = NVRW_CNT; + + memcpy(nvram_buf, p, len); + + if ((rtas_call(nvram_store, 3, 2, &done, i, __pa(nvram_buf), + len) != 0) || len != done) { + spin_unlock_irqrestore(&nvram_lock, flags); + return -EIO; + } + + p += len; + i += len; + } + spin_unlock_irqrestore(&nvram_lock, flags); + + *index = i; + return p - buf; +} + +static ssize_t pSeries_nvram_get_size(void) +{ + return nvram_size ? nvram_size : -ENODEV; +} + +/* nvram_write_error_log + * + * We need to buffer the error logs into nvram to ensure that we have + * the failure information to decode. + */ +int nvram_write_error_log(char * buff, int length, + unsigned int err_type, unsigned int error_log_cnt) +{ + int rc = nvram_write_os_partition(&rtas_log_partition, buff, length, + err_type, error_log_cnt); + if (!rc) { + last_unread_rtas_event = ktime_get_real_seconds(); +#ifdef CONFIG_PSTORE + last_rtas_event = ktime_get_real_seconds(); +#endif + } + + return rc; +} + +/* nvram_read_error_log + * + * Reads nvram for error log for at most 'length' + */ +int nvram_read_error_log(char *buff, int length, + unsigned int *err_type, unsigned int *error_log_cnt) +{ + return nvram_read_partition(&rtas_log_partition, buff, length, + err_type, error_log_cnt); +} + +/* This doesn't actually zero anything, but it sets the event_logged + * word to tell that this event is safely in syslog. + */ +int nvram_clear_error_log(void) +{ + loff_t tmp_index; + int clear_word = ERR_FLAG_ALREADY_LOGGED; + int rc; + + if (rtas_log_partition.index == -1) + return -1; + + tmp_index = rtas_log_partition.index; + + rc = ppc_md.nvram_write((char *)&clear_word, sizeof(int), &tmp_index); + if (rc <= 0) { + printk(KERN_ERR "nvram_clear_error_log: Failed nvram_write (%d)\n", rc); + return rc; + } + last_unread_rtas_event = 0; + + return 0; +} + +/* + * Are we using the ibm,rtas-log for oops/panic reports? And if so, + * would logging this oops/panic overwrite an RTAS event that rtas_errd + * hasn't had a chance to read and process? Return 1 if so, else 0. + * + * We assume that if rtas_errd hasn't read the RTAS event in + * NVRAM_RTAS_READ_TIMEOUT seconds, it's probably not going to. + */ +int clobbering_unread_rtas_event(void) +{ + return (oops_log_partition.index == rtas_log_partition.index + && last_unread_rtas_event + && ktime_get_real_seconds() - last_unread_rtas_event <= + NVRAM_RTAS_READ_TIMEOUT); +} + +static int __init pseries_nvram_init_log_partitions(void) +{ + int rc; + + /* Scan nvram for partitions */ + nvram_scan_partitions(); + + rc = nvram_init_os_partition(&rtas_log_partition); + nvram_init_oops_partition(rc == 0); + return 0; +} +machine_arch_initcall(pseries, pseries_nvram_init_log_partitions); + +int __init pSeries_nvram_init(void) +{ + struct device_node *nvram; + const __be32 *nbytes_p; + unsigned int proplen; + + nvram = of_find_node_by_type(NULL, "nvram"); + if (nvram == NULL) + return -ENODEV; + + nbytes_p = of_get_property(nvram, "#bytes", &proplen); + if (nbytes_p == NULL || proplen != sizeof(unsigned int)) { + of_node_put(nvram); + return -EIO; + } + + nvram_size = be32_to_cpup(nbytes_p); + + nvram_fetch = rtas_function_token(RTAS_FN_NVRAM_FETCH); + nvram_store = rtas_function_token(RTAS_FN_NVRAM_STORE); + printk(KERN_INFO "PPC64 nvram contains %d bytes\n", nvram_size); + of_node_put(nvram); + + ppc_md.nvram_read = pSeries_nvram_read; + ppc_md.nvram_write = pSeries_nvram_write; + ppc_md.nvram_size = pSeries_nvram_get_size; + + return 0; +} + diff --git a/arch/powerpc/platforms/pseries/of_helpers.c b/arch/powerpc/platforms/pseries/of_helpers.c new file mode 100644 index 000000000..23241c71e --- /dev/null +++ b/arch/powerpc/platforms/pseries/of_helpers.c @@ -0,0 +1,97 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <linux/string.h> +#include <linux/err.h> +#include <linux/slab.h> +#include <linux/of.h> +#include <asm/prom.h> + +#include "of_helpers.h" + +/** + * pseries_of_derive_parent - basically like dirname(1) + * @path: the full_name of a node to be added to the tree + * + * Returns the node which should be the parent of the node + * described by path. E.g., for path = "/foo/bar", returns + * the node with full_name = "/foo". + */ +struct device_node *pseries_of_derive_parent(const char *path) +{ + struct device_node *parent; + char *parent_path = "/"; + const char *tail; + + /* We do not want the trailing '/' character */ + tail = kbasename(path) - 1; + + /* reject if path is "/" */ + if (!strcmp(path, "/")) + return ERR_PTR(-EINVAL); + + if (tail > path) { + parent_path = kstrndup(path, tail - path, GFP_KERNEL); + if (!parent_path) + return ERR_PTR(-ENOMEM); + } + parent = of_find_node_by_path(parent_path); + if (strcmp(parent_path, "/")) + kfree(parent_path); + return parent ? parent : ERR_PTR(-EINVAL); +} + + +/* Helper Routines to convert between drc_index to cpu numbers */ + +int of_read_drc_info_cell(struct property **prop, const __be32 **curval, + struct of_drc_info *data) +{ + const char *p = (char *)(*curval); + const __be32 *p2; + + if (!data) + return -EINVAL; + + /* Get drc-type:encode-string */ + data->drc_type = (char *)p; + p = of_prop_next_string(*prop, p); + if (!p) + return -EINVAL; + + /* Get drc-name-prefix:encode-string */ + data->drc_name_prefix = (char *)p; + p = of_prop_next_string(*prop, p); + if (!p) + return -EINVAL; + + /* Get drc-index-start:encode-int */ + p2 = (const __be32 *)p; + data->drc_index_start = be32_to_cpu(*p2); + + /* Get drc-name-suffix-start:encode-int */ + p2 = of_prop_next_u32(*prop, p2, &data->drc_name_suffix_start); + if (!p2) + return -EINVAL; + + /* Get number-sequential-elements:encode-int */ + p2 = of_prop_next_u32(*prop, p2, &data->num_sequential_elems); + if (!p2) + return -EINVAL; + + /* Get sequential-increment:encode-int */ + p2 = of_prop_next_u32(*prop, p2, &data->sequential_inc); + if (!p2) + return -EINVAL; + + /* Get drc-power-domain:encode-int */ + p2 = of_prop_next_u32(*prop, p2, &data->drc_power_domain); + if (!p2) + return -EINVAL; + + /* Should now know end of current entry */ + (*curval) = (void *)(++p2); + data->last_drc_index = data->drc_index_start + + ((data->num_sequential_elems - 1) * data->sequential_inc); + + return 0; +} +EXPORT_SYMBOL(of_read_drc_info_cell); diff --git a/arch/powerpc/platforms/pseries/of_helpers.h b/arch/powerpc/platforms/pseries/of_helpers.h new file mode 100644 index 000000000..decad6553 --- /dev/null +++ b/arch/powerpc/platforms/pseries/of_helpers.h @@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _PSERIES_OF_HELPERS_H +#define _PSERIES_OF_HELPERS_H + +#include <linux/of.h> + +struct device_node *pseries_of_derive_parent(const char *path); + +#endif /* _PSERIES_OF_HELPERS_H */ diff --git a/arch/powerpc/platforms/pseries/papr-sysparm.c b/arch/powerpc/platforms/pseries/papr-sysparm.c new file mode 100644 index 000000000..fedc61599 --- /dev/null +++ b/arch/powerpc/platforms/pseries/papr-sysparm.c @@ -0,0 +1,151 @@ +// SPDX-License-Identifier: GPL-2.0-only + +#define pr_fmt(fmt) "papr-sysparm: " fmt + +#include <linux/bug.h> +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/printk.h> +#include <linux/slab.h> +#include <asm/rtas.h> +#include <asm/papr-sysparm.h> +#include <asm/rtas-work-area.h> + +struct papr_sysparm_buf *papr_sysparm_buf_alloc(void) +{ + struct papr_sysparm_buf *buf = kzalloc(sizeof(*buf), GFP_KERNEL); + + return buf; +} + +void papr_sysparm_buf_free(struct papr_sysparm_buf *buf) +{ + kfree(buf); +} + +/** + * papr_sysparm_get() - Retrieve the value of a PAPR system parameter. + * @param: PAPR system parameter token as described in + * 7.3.16 "System Parameters Option". + * @buf: A &struct papr_sysparm_buf as returned from papr_sysparm_buf_alloc(). + * + * Place the result of querying the specified parameter, if available, + * in @buf. The result includes a be16 length header followed by the + * value, which may be a string or binary data. See &struct papr_sysparm_buf. + * + * Since there is at least one parameter (60, OS Service Entitlement + * Status) where the results depend on the incoming contents of the + * work area, the caller-supplied buffer is copied unmodified into the + * work area before calling ibm,get-system-parameter. + * + * A defined parameter may not be implemented on a given system, and + * some implemented parameters may not be available to all partitions + * on a system. A parameter's disposition may change at any time due + * to system configuration changes or partition migration. + * + * Context: This function may sleep. + * + * Return: 0 on success, -errno otherwise. @buf is unmodified on error. + */ + +int papr_sysparm_get(papr_sysparm_t param, struct papr_sysparm_buf *buf) +{ + const s32 token = rtas_function_token(RTAS_FN_IBM_GET_SYSTEM_PARAMETER); + struct rtas_work_area *work_area; + s32 fwrc; + int ret; + + might_sleep(); + + if (WARN_ON(!buf)) + return -EFAULT; + + if (token == RTAS_UNKNOWN_SERVICE) + return -ENOENT; + + work_area = rtas_work_area_alloc(sizeof(*buf)); + + memcpy(rtas_work_area_raw_buf(work_area), buf, sizeof(*buf)); + + do { + fwrc = rtas_call(token, 3, 1, NULL, param.token, + rtas_work_area_phys(work_area), + rtas_work_area_size(work_area)); + } while (rtas_busy_delay(fwrc)); + + switch (fwrc) { + case 0: + ret = 0; + memcpy(buf, rtas_work_area_raw_buf(work_area), sizeof(*buf)); + break; + case -3: /* parameter not implemented */ + ret = -EOPNOTSUPP; + break; + case -9002: /* this partition not authorized to retrieve this parameter */ + ret = -EPERM; + break; + case -9999: /* "parameter error" e.g. the buffer is too small */ + ret = -EINVAL; + break; + default: + pr_err("unexpected ibm,get-system-parameter result %d\n", fwrc); + fallthrough; + case -1: /* Hardware/platform error */ + ret = -EIO; + break; + } + + rtas_work_area_free(work_area); + + return ret; +} + +int papr_sysparm_set(papr_sysparm_t param, const struct papr_sysparm_buf *buf) +{ + const s32 token = rtas_function_token(RTAS_FN_IBM_SET_SYSTEM_PARAMETER); + struct rtas_work_area *work_area; + s32 fwrc; + int ret; + + might_sleep(); + + if (WARN_ON(!buf)) + return -EFAULT; + + if (token == RTAS_UNKNOWN_SERVICE) + return -ENOENT; + + work_area = rtas_work_area_alloc(sizeof(*buf)); + + memcpy(rtas_work_area_raw_buf(work_area), buf, sizeof(*buf)); + + do { + fwrc = rtas_call(token, 2, 1, NULL, param.token, + rtas_work_area_phys(work_area)); + } while (rtas_busy_delay(fwrc)); + + switch (fwrc) { + case 0: + ret = 0; + break; + case -3: /* parameter not supported */ + ret = -EOPNOTSUPP; + break; + case -9002: /* this partition not authorized to modify this parameter */ + ret = -EPERM; + break; + case -9999: /* "parameter error" e.g. invalid input data */ + ret = -EINVAL; + break; + default: + pr_err("unexpected ibm,set-system-parameter result %d\n", fwrc); + fallthrough; + case -1: /* Hardware/platform error */ + ret = -EIO; + break; + } + + rtas_work_area_free(work_area); + + return ret; +} diff --git a/arch/powerpc/platforms/pseries/papr_platform_attributes.c b/arch/powerpc/platforms/pseries/papr_platform_attributes.c new file mode 100644 index 000000000..526c621b0 --- /dev/null +++ b/arch/powerpc/platforms/pseries/papr_platform_attributes.c @@ -0,0 +1,362 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Platform energy and frequency attributes driver + * + * This driver creates a sys file at /sys/firmware/papr/ which encapsulates a + * directory structure containing files in keyword - value pairs that specify + * energy and frequency configuration of the system. + * + * The format of exposing the sysfs information is as follows: + * /sys/firmware/papr/energy_scale_info/ + * |-- <id>/ + * |-- desc + * |-- value + * |-- value_desc (if exists) + * |-- <id>/ + * |-- desc + * |-- value + * |-- value_desc (if exists) + * + * Copyright 2022 IBM Corp. + */ + +#include <asm/hvcall.h> +#include <asm/machdep.h> +#include <asm/firmware.h> + +#include "pseries.h" + +/* + * Flag attributes to fetch either all or one attribute from the HCALL + * flag = BE(0) => fetch all attributes with firstAttributeId = 0 + * flag = BE(1) => fetch a single attribute with firstAttributeId = id + */ +#define ESI_FLAGS_ALL 0 +#define ESI_FLAGS_SINGLE (1ull << 63) + +#define KOBJ_MAX_ATTRS 3 + +#define ESI_HDR_SIZE sizeof(struct h_energy_scale_info_hdr) +#define ESI_ATTR_SIZE sizeof(struct energy_scale_attribute) +#define CURR_MAX_ESI_ATTRS 8 + +struct energy_scale_attribute { + __be64 id; + __be64 val; + u8 desc[64]; + u8 value_desc[64]; +} __packed; + +struct h_energy_scale_info_hdr { + __be64 num_attrs; + __be64 array_offset; + u8 data_header_version; +} __packed; + +struct papr_attr { + u64 id; + struct kobj_attribute kobj_attr; +}; + +struct papr_group { + struct attribute_group pg; + struct papr_attr pgattrs[KOBJ_MAX_ATTRS]; +}; + +static struct papr_group *papr_groups; +/* /sys/firmware/papr */ +static struct kobject *papr_kobj; +/* /sys/firmware/papr/energy_scale_info */ +static struct kobject *esi_kobj; + +/* + * Energy modes can change dynamically hence making a new hcall each time the + * information needs to be retrieved + */ +static int papr_get_attr(u64 id, struct energy_scale_attribute *esi) +{ + int esi_buf_size = ESI_HDR_SIZE + (CURR_MAX_ESI_ATTRS * ESI_ATTR_SIZE); + int ret, max_esi_attrs = CURR_MAX_ESI_ATTRS; + struct energy_scale_attribute *curr_esi; + struct h_energy_scale_info_hdr *hdr; + char *buf; + + buf = kmalloc(esi_buf_size, GFP_KERNEL); + if (buf == NULL) + return -ENOMEM; + +retry: + ret = plpar_hcall_norets(H_GET_ENERGY_SCALE_INFO, ESI_FLAGS_SINGLE, + id, virt_to_phys(buf), + esi_buf_size); + + /* + * If the hcall fails with not enough memory for either the + * header or data, attempt to allocate more + */ + if (ret == H_PARTIAL || ret == H_P4) { + char *temp_buf; + + max_esi_attrs += 4; + esi_buf_size = ESI_HDR_SIZE + (CURR_MAX_ESI_ATTRS * max_esi_attrs); + + temp_buf = krealloc(buf, esi_buf_size, GFP_KERNEL); + if (temp_buf) + buf = temp_buf; + else + return -ENOMEM; + + goto retry; + } + + if (ret != H_SUCCESS) { + pr_warn("hcall failed: H_GET_ENERGY_SCALE_INFO"); + ret = -EIO; + goto out_buf; + } + + hdr = (struct h_energy_scale_info_hdr *) buf; + curr_esi = (struct energy_scale_attribute *) + (buf + be64_to_cpu(hdr->array_offset)); + + if (esi_buf_size < + be64_to_cpu(hdr->array_offset) + (be64_to_cpu(hdr->num_attrs) + * sizeof(struct energy_scale_attribute))) { + ret = -EIO; + goto out_buf; + } + + *esi = *curr_esi; + +out_buf: + kfree(buf); + + return ret; +} + +/* + * Extract and export the description of the energy scale attributes + */ +static ssize_t desc_show(struct kobject *kobj, + struct kobj_attribute *kobj_attr, + char *buf) +{ + struct papr_attr *pattr = container_of(kobj_attr, struct papr_attr, + kobj_attr); + struct energy_scale_attribute esi; + int ret; + + ret = papr_get_attr(pattr->id, &esi); + if (ret) + return ret; + + return sysfs_emit(buf, "%s\n", esi.desc); +} + +/* + * Extract and export the numeric value of the energy scale attributes + */ +static ssize_t val_show(struct kobject *kobj, + struct kobj_attribute *kobj_attr, + char *buf) +{ + struct papr_attr *pattr = container_of(kobj_attr, struct papr_attr, + kobj_attr); + struct energy_scale_attribute esi; + int ret; + + ret = papr_get_attr(pattr->id, &esi); + if (ret) + return ret; + + return sysfs_emit(buf, "%llu\n", be64_to_cpu(esi.val)); +} + +/* + * Extract and export the value description in string format of the energy + * scale attributes + */ +static ssize_t val_desc_show(struct kobject *kobj, + struct kobj_attribute *kobj_attr, + char *buf) +{ + struct papr_attr *pattr = container_of(kobj_attr, struct papr_attr, + kobj_attr); + struct energy_scale_attribute esi; + int ret; + + ret = papr_get_attr(pattr->id, &esi); + if (ret) + return ret; + + return sysfs_emit(buf, "%s\n", esi.value_desc); +} + +static struct papr_ops_info { + const char *attr_name; + ssize_t (*show)(struct kobject *kobj, struct kobj_attribute *kobj_attr, + char *buf); +} ops_info[KOBJ_MAX_ATTRS] = { + { "desc", desc_show }, + { "value", val_show }, + { "value_desc", val_desc_show }, +}; + +static void add_attr(u64 id, int index, struct papr_attr *attr) +{ + attr->id = id; + sysfs_attr_init(&attr->kobj_attr.attr); + attr->kobj_attr.attr.name = ops_info[index].attr_name; + attr->kobj_attr.attr.mode = 0444; + attr->kobj_attr.show = ops_info[index].show; +} + +static int add_attr_group(u64 id, struct papr_group *pg, bool show_val_desc) +{ + int i; + + for (i = 0; i < KOBJ_MAX_ATTRS; i++) { + if (!strcmp(ops_info[i].attr_name, "value_desc") && + !show_val_desc) { + continue; + } + add_attr(id, i, &pg->pgattrs[i]); + pg->pg.attrs[i] = &pg->pgattrs[i].kobj_attr.attr; + } + + return sysfs_create_group(esi_kobj, &pg->pg); +} + + +static int __init papr_init(void) +{ + int esi_buf_size = ESI_HDR_SIZE + (CURR_MAX_ESI_ATTRS * ESI_ATTR_SIZE); + int ret, idx, i, max_esi_attrs = CURR_MAX_ESI_ATTRS; + struct h_energy_scale_info_hdr *esi_hdr; + struct energy_scale_attribute *esi_attrs; + uint64_t num_attrs; + char *esi_buf; + + if (!firmware_has_feature(FW_FEATURE_LPAR) || + !firmware_has_feature(FW_FEATURE_ENERGY_SCALE_INFO)) { + return -ENXIO; + } + + esi_buf = kmalloc(esi_buf_size, GFP_KERNEL); + if (esi_buf == NULL) + return -ENOMEM; + /* + * hcall( + * uint64 H_GET_ENERGY_SCALE_INFO, // Get energy scale info + * uint64 flags, // Per the flag request + * uint64 firstAttributeId, // The attribute id + * uint64 bufferAddress, // Guest physical address of the output buffer + * uint64 bufferSize); // The size in bytes of the output buffer + */ +retry: + + ret = plpar_hcall_norets(H_GET_ENERGY_SCALE_INFO, ESI_FLAGS_ALL, 0, + virt_to_phys(esi_buf), esi_buf_size); + + /* + * If the hcall fails with not enough memory for either the + * header or data, attempt to allocate more + */ + if (ret == H_PARTIAL || ret == H_P4) { + char *temp_esi_buf; + + max_esi_attrs += 4; + esi_buf_size = ESI_HDR_SIZE + (CURR_MAX_ESI_ATTRS * max_esi_attrs); + + temp_esi_buf = krealloc(esi_buf, esi_buf_size, GFP_KERNEL); + if (temp_esi_buf) + esi_buf = temp_esi_buf; + else + return -ENOMEM; + + goto retry; + } + + if (ret != H_SUCCESS) { + pr_warn("hcall failed: H_GET_ENERGY_SCALE_INFO, ret: %d\n", ret); + goto out_free_esi_buf; + } + + esi_hdr = (struct h_energy_scale_info_hdr *) esi_buf; + num_attrs = be64_to_cpu(esi_hdr->num_attrs); + esi_attrs = (struct energy_scale_attribute *) + (esi_buf + be64_to_cpu(esi_hdr->array_offset)); + + if (esi_buf_size < + be64_to_cpu(esi_hdr->array_offset) + + (num_attrs * sizeof(struct energy_scale_attribute))) { + goto out_free_esi_buf; + } + + papr_groups = kcalloc(num_attrs, sizeof(*papr_groups), GFP_KERNEL); + if (!papr_groups) + goto out_free_esi_buf; + + papr_kobj = kobject_create_and_add("papr", firmware_kobj); + if (!papr_kobj) { + pr_warn("kobject_create_and_add papr failed\n"); + goto out_papr_groups; + } + + esi_kobj = kobject_create_and_add("energy_scale_info", papr_kobj); + if (!esi_kobj) { + pr_warn("kobject_create_and_add energy_scale_info failed\n"); + goto out_kobj; + } + + /* Allocate the groups before registering */ + for (idx = 0; idx < num_attrs; idx++) { + papr_groups[idx].pg.attrs = kcalloc(KOBJ_MAX_ATTRS + 1, + sizeof(*papr_groups[idx].pg.attrs), + GFP_KERNEL); + if (!papr_groups[idx].pg.attrs) + goto out_pgattrs; + + papr_groups[idx].pg.name = kasprintf(GFP_KERNEL, "%lld", + be64_to_cpu(esi_attrs[idx].id)); + if (papr_groups[idx].pg.name == NULL) + goto out_pgattrs; + } + + for (idx = 0; idx < num_attrs; idx++) { + bool show_val_desc = true; + + /* Do not add the value desc attr if it does not exist */ + if (strnlen(esi_attrs[idx].value_desc, + sizeof(esi_attrs[idx].value_desc)) == 0) + show_val_desc = false; + + if (add_attr_group(be64_to_cpu(esi_attrs[idx].id), + &papr_groups[idx], + show_val_desc)) { + pr_warn("Failed to create papr attribute group %s\n", + papr_groups[idx].pg.name); + idx = num_attrs; + goto out_pgattrs; + } + } + + kfree(esi_buf); + return 0; +out_pgattrs: + for (i = 0; i < idx ; i++) { + kfree(papr_groups[i].pg.attrs); + kfree(papr_groups[i].pg.name); + } + kobject_put(esi_kobj); +out_kobj: + kobject_put(papr_kobj); +out_papr_groups: + kfree(papr_groups); +out_free_esi_buf: + kfree(esi_buf); + + return -ENOMEM; +} + +machine_device_initcall(pseries, papr_init); diff --git a/arch/powerpc/platforms/pseries/papr_scm.c b/arch/powerpc/platforms/pseries/papr_scm.c new file mode 100644 index 000000000..1a53e048c --- /dev/null +++ b/arch/powerpc/platforms/pseries/papr_scm.c @@ -0,0 +1,1581 @@ +// SPDX-License-Identifier: GPL-2.0 + +#define pr_fmt(fmt) "papr-scm: " fmt + +#include <linux/of.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/ioport.h> +#include <linux/slab.h> +#include <linux/ndctl.h> +#include <linux/sched.h> +#include <linux/libnvdimm.h> +#include <linux/platform_device.h> +#include <linux/delay.h> +#include <linux/seq_buf.h> +#include <linux/nd.h> + +#include <asm/plpar_wrappers.h> +#include <asm/papr_pdsm.h> +#include <asm/mce.h> +#include <asm/unaligned.h> +#include <linux/perf_event.h> + +#define BIND_ANY_ADDR (~0ul) + +#define PAPR_SCM_DIMM_CMD_MASK \ + ((1ul << ND_CMD_GET_CONFIG_SIZE) | \ + (1ul << ND_CMD_GET_CONFIG_DATA) | \ + (1ul << ND_CMD_SET_CONFIG_DATA) | \ + (1ul << ND_CMD_CALL)) + +/* DIMM health bitmap indicators */ +/* SCM device is unable to persist memory contents */ +#define PAPR_PMEM_UNARMED (1ULL << (63 - 0)) +/* SCM device failed to persist memory contents */ +#define PAPR_PMEM_SHUTDOWN_DIRTY (1ULL << (63 - 1)) +/* SCM device contents are persisted from previous IPL */ +#define PAPR_PMEM_SHUTDOWN_CLEAN (1ULL << (63 - 2)) +/* SCM device contents are not persisted from previous IPL */ +#define PAPR_PMEM_EMPTY (1ULL << (63 - 3)) +/* SCM device memory life remaining is critically low */ +#define PAPR_PMEM_HEALTH_CRITICAL (1ULL << (63 - 4)) +/* SCM device will be garded off next IPL due to failure */ +#define PAPR_PMEM_HEALTH_FATAL (1ULL << (63 - 5)) +/* SCM contents cannot persist due to current platform health status */ +#define PAPR_PMEM_HEALTH_UNHEALTHY (1ULL << (63 - 6)) +/* SCM device is unable to persist memory contents in certain conditions */ +#define PAPR_PMEM_HEALTH_NON_CRITICAL (1ULL << (63 - 7)) +/* SCM device is encrypted */ +#define PAPR_PMEM_ENCRYPTED (1ULL << (63 - 8)) +/* SCM device has been scrubbed and locked */ +#define PAPR_PMEM_SCRUBBED_AND_LOCKED (1ULL << (63 - 9)) + +/* Bits status indicators for health bitmap indicating unarmed dimm */ +#define PAPR_PMEM_UNARMED_MASK (PAPR_PMEM_UNARMED | \ + PAPR_PMEM_HEALTH_UNHEALTHY) + +/* Bits status indicators for health bitmap indicating unflushed dimm */ +#define PAPR_PMEM_BAD_SHUTDOWN_MASK (PAPR_PMEM_SHUTDOWN_DIRTY) + +/* Bits status indicators for health bitmap indicating unrestored dimm */ +#define PAPR_PMEM_BAD_RESTORE_MASK (PAPR_PMEM_EMPTY) + +/* Bit status indicators for smart event notification */ +#define PAPR_PMEM_SMART_EVENT_MASK (PAPR_PMEM_HEALTH_CRITICAL | \ + PAPR_PMEM_HEALTH_FATAL | \ + PAPR_PMEM_HEALTH_UNHEALTHY) + +#define PAPR_SCM_PERF_STATS_EYECATCHER __stringify(SCMSTATS) +#define PAPR_SCM_PERF_STATS_VERSION 0x1 + +/* Struct holding a single performance metric */ +struct papr_scm_perf_stat { + u8 stat_id[8]; + __be64 stat_val; +} __packed; + +/* Struct exchanged between kernel and PHYP for fetching drc perf stats */ +struct papr_scm_perf_stats { + u8 eye_catcher[8]; + /* Should be PAPR_SCM_PERF_STATS_VERSION */ + __be32 stats_version; + /* Number of stats following */ + __be32 num_statistics; + /* zero or more performance matrics */ + struct papr_scm_perf_stat scm_statistic[]; +} __packed; + +/* private struct associated with each region */ +struct papr_scm_priv { + struct platform_device *pdev; + struct device_node *dn; + uint32_t drc_index; + uint64_t blocks; + uint64_t block_size; + int metadata_size; + bool is_volatile; + bool hcall_flush_required; + + uint64_t bound_addr; + + struct nvdimm_bus_descriptor bus_desc; + struct nvdimm_bus *bus; + struct nvdimm *nvdimm; + struct resource res; + struct nd_region *region; + struct nd_interleave_set nd_set; + struct list_head region_list; + + /* Protect dimm health data from concurrent read/writes */ + struct mutex health_mutex; + + /* Last time the health information of the dimm was updated */ + unsigned long lasthealth_jiffies; + + /* Health information for the dimm */ + u64 health_bitmap; + + /* Holds the last known dirty shutdown counter value */ + u64 dirty_shutdown_counter; + + /* length of the stat buffer as expected by phyp */ + size_t stat_buffer_len; + + /* The bits which needs to be overridden */ + u64 health_bitmap_inject_mask; +}; + +static int papr_scm_pmem_flush(struct nd_region *nd_region, + struct bio *bio __maybe_unused) +{ + struct papr_scm_priv *p = nd_region_provider_data(nd_region); + unsigned long ret_buf[PLPAR_HCALL_BUFSIZE], token = 0; + long rc; + + dev_dbg(&p->pdev->dev, "flush drc 0x%x", p->drc_index); + + do { + rc = plpar_hcall(H_SCM_FLUSH, ret_buf, p->drc_index, token); + token = ret_buf[0]; + + /* Check if we are stalled for some time */ + if (H_IS_LONG_BUSY(rc)) { + msleep(get_longbusy_msecs(rc)); + rc = H_BUSY; + } else if (rc == H_BUSY) { + cond_resched(); + } + } while (rc == H_BUSY); + + if (rc) { + dev_err(&p->pdev->dev, "flush error: %ld", rc); + rc = -EIO; + } else { + dev_dbg(&p->pdev->dev, "flush drc 0x%x complete", p->drc_index); + } + + return rc; +} + +static LIST_HEAD(papr_nd_regions); +static DEFINE_MUTEX(papr_ndr_lock); + +static int drc_pmem_bind(struct papr_scm_priv *p) +{ + unsigned long ret[PLPAR_HCALL_BUFSIZE]; + uint64_t saved = 0; + uint64_t token; + int64_t rc; + + /* + * When the hypervisor cannot map all the requested memory in a single + * hcall it returns H_BUSY and we call again with the token until + * we get H_SUCCESS. Aborting the retry loop before getting H_SUCCESS + * leave the system in an undefined state, so we wait. + */ + token = 0; + + do { + rc = plpar_hcall(H_SCM_BIND_MEM, ret, p->drc_index, 0, + p->blocks, BIND_ANY_ADDR, token); + token = ret[0]; + if (!saved) + saved = ret[1]; + cond_resched(); + } while (rc == H_BUSY); + + if (rc) + return rc; + + p->bound_addr = saved; + dev_dbg(&p->pdev->dev, "bound drc 0x%x to 0x%lx\n", + p->drc_index, (unsigned long)saved); + return rc; +} + +static void drc_pmem_unbind(struct papr_scm_priv *p) +{ + unsigned long ret[PLPAR_HCALL_BUFSIZE]; + uint64_t token = 0; + int64_t rc; + + dev_dbg(&p->pdev->dev, "unbind drc 0x%x\n", p->drc_index); + + /* NB: unbind has the same retry requirements as drc_pmem_bind() */ + do { + + /* Unbind of all SCM resources associated with drcIndex */ + rc = plpar_hcall(H_SCM_UNBIND_ALL, ret, H_UNBIND_SCOPE_DRC, + p->drc_index, token); + token = ret[0]; + + /* Check if we are stalled for some time */ + if (H_IS_LONG_BUSY(rc)) { + msleep(get_longbusy_msecs(rc)); + rc = H_BUSY; + } else if (rc == H_BUSY) { + cond_resched(); + } + + } while (rc == H_BUSY); + + if (rc) + dev_err(&p->pdev->dev, "unbind error: %lld\n", rc); + else + dev_dbg(&p->pdev->dev, "unbind drc 0x%x complete\n", + p->drc_index); + + return; +} + +static int drc_pmem_query_n_bind(struct papr_scm_priv *p) +{ + unsigned long start_addr; + unsigned long end_addr; + unsigned long ret[PLPAR_HCALL_BUFSIZE]; + int64_t rc; + + + rc = plpar_hcall(H_SCM_QUERY_BLOCK_MEM_BINDING, ret, + p->drc_index, 0); + if (rc) + goto err_out; + start_addr = ret[0]; + + /* Make sure the full region is bound. */ + rc = plpar_hcall(H_SCM_QUERY_BLOCK_MEM_BINDING, ret, + p->drc_index, p->blocks - 1); + if (rc) + goto err_out; + end_addr = ret[0]; + + if ((end_addr - start_addr) != ((p->blocks - 1) * p->block_size)) + goto err_out; + + p->bound_addr = start_addr; + dev_dbg(&p->pdev->dev, "bound drc 0x%x to 0x%lx\n", p->drc_index, start_addr); + return rc; + +err_out: + dev_info(&p->pdev->dev, + "Failed to query, trying an unbind followed by bind"); + drc_pmem_unbind(p); + return drc_pmem_bind(p); +} + +/* + * Query the Dimm performance stats from PHYP and copy them (if returned) to + * provided struct papr_scm_perf_stats instance 'stats' that can hold atleast + * (num_stats + header) bytes. + * - If buff_stats == NULL the return value is the size in bytes of the buffer + * needed to hold all supported performance-statistics. + * - If buff_stats != NULL and num_stats == 0 then we copy all known + * performance-statistics to 'buff_stat' and expect to be large enough to + * hold them. + * - if buff_stats != NULL and num_stats > 0 then copy the requested + * performance-statistics to buff_stats. + */ +static ssize_t drc_pmem_query_stats(struct papr_scm_priv *p, + struct papr_scm_perf_stats *buff_stats, + unsigned int num_stats) +{ + unsigned long ret[PLPAR_HCALL_BUFSIZE]; + size_t size; + s64 rc; + + /* Setup the out buffer */ + if (buff_stats) { + memcpy(buff_stats->eye_catcher, + PAPR_SCM_PERF_STATS_EYECATCHER, 8); + buff_stats->stats_version = + cpu_to_be32(PAPR_SCM_PERF_STATS_VERSION); + buff_stats->num_statistics = + cpu_to_be32(num_stats); + + /* + * Calculate the buffer size based on num-stats provided + * or use the prefetched max buffer length + */ + if (num_stats) + /* Calculate size from the num_stats */ + size = sizeof(struct papr_scm_perf_stats) + + num_stats * sizeof(struct papr_scm_perf_stat); + else + size = p->stat_buffer_len; + } else { + /* In case of no out buffer ignore the size */ + size = 0; + } + + /* Do the HCALL asking PHYP for info */ + rc = plpar_hcall(H_SCM_PERFORMANCE_STATS, ret, p->drc_index, + buff_stats ? virt_to_phys(buff_stats) : 0, + size); + + /* Check if the error was due to an unknown stat-id */ + if (rc == H_PARTIAL) { + dev_err(&p->pdev->dev, + "Unknown performance stats, Err:0x%016lX\n", ret[0]); + return -ENOENT; + } else if (rc == H_AUTHORITY) { + dev_info(&p->pdev->dev, + "Permission denied while accessing performance stats"); + return -EPERM; + } else if (rc == H_UNSUPPORTED) { + dev_dbg(&p->pdev->dev, "Performance stats unsupported\n"); + return -EOPNOTSUPP; + } else if (rc != H_SUCCESS) { + dev_err(&p->pdev->dev, + "Failed to query performance stats, Err:%lld\n", rc); + return -EIO; + + } else if (!size) { + /* Handle case where stat buffer size was requested */ + dev_dbg(&p->pdev->dev, + "Performance stats size %ld\n", ret[0]); + return ret[0]; + } + + /* Successfully fetched the requested stats from phyp */ + dev_dbg(&p->pdev->dev, + "Performance stats returned %d stats\n", + be32_to_cpu(buff_stats->num_statistics)); + return 0; +} + +#ifdef CONFIG_PERF_EVENTS +#define to_nvdimm_pmu(_pmu) container_of(_pmu, struct nvdimm_pmu, pmu) + +static const char * const nvdimm_events_map[] = { + [1] = "CtlResCt", + [2] = "CtlResTm", + [3] = "PonSecs ", + [4] = "MemLife ", + [5] = "CritRscU", + [6] = "HostLCnt", + [7] = "HostSCnt", + [8] = "HostSDur", + [9] = "HostLDur", + [10] = "MedRCnt ", + [11] = "MedWCnt ", + [12] = "MedRDur ", + [13] = "MedWDur ", + [14] = "CchRHCnt", + [15] = "CchWHCnt", + [16] = "FastWCnt", +}; + +static int papr_scm_pmu_get_value(struct perf_event *event, struct device *dev, u64 *count) +{ + struct papr_scm_perf_stat *stat; + struct papr_scm_perf_stats *stats; + struct papr_scm_priv *p = dev_get_drvdata(dev); + int rc, size; + + /* Invalid eventcode */ + if (event->attr.config == 0 || event->attr.config >= ARRAY_SIZE(nvdimm_events_map)) + return -EINVAL; + + /* Allocate request buffer enough to hold single performance stat */ + size = sizeof(struct papr_scm_perf_stats) + + sizeof(struct papr_scm_perf_stat); + + if (!p) + return -EINVAL; + + stats = kzalloc(size, GFP_KERNEL); + if (!stats) + return -ENOMEM; + + stat = &stats->scm_statistic[0]; + memcpy(&stat->stat_id, + nvdimm_events_map[event->attr.config], + sizeof(stat->stat_id)); + stat->stat_val = 0; + + rc = drc_pmem_query_stats(p, stats, 1); + if (rc < 0) { + kfree(stats); + return rc; + } + + *count = be64_to_cpu(stat->stat_val); + kfree(stats); + return 0; +} + +static int papr_scm_pmu_event_init(struct perf_event *event) +{ + struct nvdimm_pmu *nd_pmu = to_nvdimm_pmu(event->pmu); + struct papr_scm_priv *p; + + if (!nd_pmu) + return -EINVAL; + + /* test the event attr type for PMU enumeration */ + if (event->attr.type != event->pmu->type) + return -ENOENT; + + /* it does not support event sampling mode */ + if (is_sampling_event(event)) + return -EOPNOTSUPP; + + /* no branch sampling */ + if (has_branch_stack(event)) + return -EOPNOTSUPP; + + p = (struct papr_scm_priv *)nd_pmu->dev->driver_data; + if (!p) + return -EINVAL; + + /* Invalid eventcode */ + if (event->attr.config == 0 || event->attr.config > 16) + return -EINVAL; + + return 0; +} + +static int papr_scm_pmu_add(struct perf_event *event, int flags) +{ + u64 count; + int rc; + struct nvdimm_pmu *nd_pmu = to_nvdimm_pmu(event->pmu); + + if (!nd_pmu) + return -EINVAL; + + if (flags & PERF_EF_START) { + rc = papr_scm_pmu_get_value(event, nd_pmu->dev, &count); + if (rc) + return rc; + + local64_set(&event->hw.prev_count, count); + } + + return 0; +} + +static void papr_scm_pmu_read(struct perf_event *event) +{ + u64 prev, now; + int rc; + struct nvdimm_pmu *nd_pmu = to_nvdimm_pmu(event->pmu); + + if (!nd_pmu) + return; + + rc = papr_scm_pmu_get_value(event, nd_pmu->dev, &now); + if (rc) + return; + + prev = local64_xchg(&event->hw.prev_count, now); + local64_add(now - prev, &event->count); +} + +static void papr_scm_pmu_del(struct perf_event *event, int flags) +{ + papr_scm_pmu_read(event); +} + +static void papr_scm_pmu_register(struct papr_scm_priv *p) +{ + struct nvdimm_pmu *nd_pmu; + int rc, nodeid; + + nd_pmu = kzalloc(sizeof(*nd_pmu), GFP_KERNEL); + if (!nd_pmu) { + rc = -ENOMEM; + goto pmu_err_print; + } + + if (!p->stat_buffer_len) { + rc = -ENOENT; + goto pmu_check_events_err; + } + + nd_pmu->pmu.task_ctx_nr = perf_invalid_context; + nd_pmu->pmu.name = nvdimm_name(p->nvdimm); + nd_pmu->pmu.event_init = papr_scm_pmu_event_init; + nd_pmu->pmu.read = papr_scm_pmu_read; + nd_pmu->pmu.add = papr_scm_pmu_add; + nd_pmu->pmu.del = papr_scm_pmu_del; + + nd_pmu->pmu.capabilities = PERF_PMU_CAP_NO_INTERRUPT | + PERF_PMU_CAP_NO_EXCLUDE; + + /*updating the cpumask variable */ + nodeid = numa_map_to_online_node(dev_to_node(&p->pdev->dev)); + nd_pmu->arch_cpumask = *cpumask_of_node(nodeid); + + rc = register_nvdimm_pmu(nd_pmu, p->pdev); + if (rc) + goto pmu_check_events_err; + + /* + * Set archdata.priv value to nvdimm_pmu structure, to handle the + * unregistering of pmu device. + */ + p->pdev->archdata.priv = nd_pmu; + return; + +pmu_check_events_err: + kfree(nd_pmu); +pmu_err_print: + dev_info(&p->pdev->dev, "nvdimm pmu didn't register rc=%d\n", rc); +} + +#else +static void papr_scm_pmu_register(struct papr_scm_priv *p) { } +#endif + +/* + * Issue hcall to retrieve dimm health info and populate papr_scm_priv with the + * health information. + */ +static int __drc_pmem_query_health(struct papr_scm_priv *p) +{ + unsigned long ret[PLPAR_HCALL_BUFSIZE]; + u64 bitmap = 0; + long rc; + + /* issue the hcall */ + rc = plpar_hcall(H_SCM_HEALTH, ret, p->drc_index); + if (rc == H_SUCCESS) + bitmap = ret[0] & ret[1]; + else if (rc == H_FUNCTION) + dev_info_once(&p->pdev->dev, + "Hcall H_SCM_HEALTH not implemented, assuming empty health bitmap"); + else { + + dev_err(&p->pdev->dev, + "Failed to query health information, Err:%ld\n", rc); + return -ENXIO; + } + + p->lasthealth_jiffies = jiffies; + /* Allow injecting specific health bits via inject mask. */ + if (p->health_bitmap_inject_mask) + bitmap = (bitmap & ~p->health_bitmap_inject_mask) | + p->health_bitmap_inject_mask; + WRITE_ONCE(p->health_bitmap, bitmap); + dev_dbg(&p->pdev->dev, + "Queried dimm health info. Bitmap:0x%016lx Mask:0x%016lx\n", + ret[0], ret[1]); + + return 0; +} + +/* Min interval in seconds for assuming stable dimm health */ +#define MIN_HEALTH_QUERY_INTERVAL 60 + +/* Query cached health info and if needed call drc_pmem_query_health */ +static int drc_pmem_query_health(struct papr_scm_priv *p) +{ + unsigned long cache_timeout; + int rc; + + /* Protect concurrent modifications to papr_scm_priv */ + rc = mutex_lock_interruptible(&p->health_mutex); + if (rc) + return rc; + + /* Jiffies offset for which the health data is assumed to be same */ + cache_timeout = p->lasthealth_jiffies + + msecs_to_jiffies(MIN_HEALTH_QUERY_INTERVAL * 1000); + + /* Fetch new health info is its older than MIN_HEALTH_QUERY_INTERVAL */ + if (time_after(jiffies, cache_timeout)) + rc = __drc_pmem_query_health(p); + else + /* Assume cached health data is valid */ + rc = 0; + + mutex_unlock(&p->health_mutex); + return rc; +} + +static int papr_scm_meta_get(struct papr_scm_priv *p, + struct nd_cmd_get_config_data_hdr *hdr) +{ + unsigned long data[PLPAR_HCALL_BUFSIZE]; + unsigned long offset, data_offset; + int len, read; + int64_t ret; + + if ((hdr->in_offset + hdr->in_length) > p->metadata_size) + return -EINVAL; + + for (len = hdr->in_length; len; len -= read) { + + data_offset = hdr->in_length - len; + offset = hdr->in_offset + data_offset; + + if (len >= 8) + read = 8; + else if (len >= 4) + read = 4; + else if (len >= 2) + read = 2; + else + read = 1; + + ret = plpar_hcall(H_SCM_READ_METADATA, data, p->drc_index, + offset, read); + + if (ret == H_PARAMETER) /* bad DRC index */ + return -ENODEV; + if (ret) + return -EINVAL; /* other invalid parameter */ + + switch (read) { + case 8: + *(uint64_t *)(hdr->out_buf + data_offset) = be64_to_cpu(data[0]); + break; + case 4: + *(uint32_t *)(hdr->out_buf + data_offset) = be32_to_cpu(data[0] & 0xffffffff); + break; + + case 2: + *(uint16_t *)(hdr->out_buf + data_offset) = be16_to_cpu(data[0] & 0xffff); + break; + + case 1: + *(uint8_t *)(hdr->out_buf + data_offset) = (data[0] & 0xff); + break; + } + } + return 0; +} + +static int papr_scm_meta_set(struct papr_scm_priv *p, + struct nd_cmd_set_config_hdr *hdr) +{ + unsigned long offset, data_offset; + int len, wrote; + unsigned long data; + __be64 data_be; + int64_t ret; + + if ((hdr->in_offset + hdr->in_length) > p->metadata_size) + return -EINVAL; + + for (len = hdr->in_length; len; len -= wrote) { + + data_offset = hdr->in_length - len; + offset = hdr->in_offset + data_offset; + + if (len >= 8) { + data = *(uint64_t *)(hdr->in_buf + data_offset); + data_be = cpu_to_be64(data); + wrote = 8; + } else if (len >= 4) { + data = *(uint32_t *)(hdr->in_buf + data_offset); + data &= 0xffffffff; + data_be = cpu_to_be32(data); + wrote = 4; + } else if (len >= 2) { + data = *(uint16_t *)(hdr->in_buf + data_offset); + data &= 0xffff; + data_be = cpu_to_be16(data); + wrote = 2; + } else { + data_be = *(uint8_t *)(hdr->in_buf + data_offset); + data_be &= 0xff; + wrote = 1; + } + + ret = plpar_hcall_norets(H_SCM_WRITE_METADATA, p->drc_index, + offset, data_be, wrote); + if (ret == H_PARAMETER) /* bad DRC index */ + return -ENODEV; + if (ret) + return -EINVAL; /* other invalid parameter */ + } + + return 0; +} + +/* + * Do a sanity checks on the inputs args to dimm-control function and return + * '0' if valid. Validation of PDSM payloads happens later in + * papr_scm_service_pdsm. + */ +static int is_cmd_valid(struct nvdimm *nvdimm, unsigned int cmd, void *buf, + unsigned int buf_len) +{ + unsigned long cmd_mask = PAPR_SCM_DIMM_CMD_MASK; + struct nd_cmd_pkg *nd_cmd; + struct papr_scm_priv *p; + enum papr_pdsm pdsm; + + /* Only dimm-specific calls are supported atm */ + if (!nvdimm) + return -EINVAL; + + /* get the provider data from struct nvdimm */ + p = nvdimm_provider_data(nvdimm); + + if (!test_bit(cmd, &cmd_mask)) { + dev_dbg(&p->pdev->dev, "Unsupported cmd=%u\n", cmd); + return -EINVAL; + } + + /* For CMD_CALL verify pdsm request */ + if (cmd == ND_CMD_CALL) { + /* Verify the envelope and envelop size */ + if (!buf || + buf_len < (sizeof(struct nd_cmd_pkg) + ND_PDSM_HDR_SIZE)) { + dev_dbg(&p->pdev->dev, "Invalid pkg size=%u\n", + buf_len); + return -EINVAL; + } + + /* Verify that the nd_cmd_pkg.nd_family is correct */ + nd_cmd = (struct nd_cmd_pkg *)buf; + + if (nd_cmd->nd_family != NVDIMM_FAMILY_PAPR) { + dev_dbg(&p->pdev->dev, "Invalid pkg family=0x%llx\n", + nd_cmd->nd_family); + return -EINVAL; + } + + pdsm = (enum papr_pdsm)nd_cmd->nd_command; + + /* Verify if the pdsm command is valid */ + if (pdsm <= PAPR_PDSM_MIN || pdsm >= PAPR_PDSM_MAX) { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Invalid PDSM\n", + pdsm); + return -EINVAL; + } + + /* Have enough space to hold returned 'nd_pkg_pdsm' header */ + if (nd_cmd->nd_size_out < ND_PDSM_HDR_SIZE) { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Invalid payload\n", + pdsm); + return -EINVAL; + } + } + + /* Let the command be further processed */ + return 0; +} + +static int papr_pdsm_fuel_gauge(struct papr_scm_priv *p, + union nd_pdsm_payload *payload) +{ + int rc, size; + u64 statval; + struct papr_scm_perf_stat *stat; + struct papr_scm_perf_stats *stats; + + /* Silently fail if fetching performance metrics isn't supported */ + if (!p->stat_buffer_len) + return 0; + + /* Allocate request buffer enough to hold single performance stat */ + size = sizeof(struct papr_scm_perf_stats) + + sizeof(struct papr_scm_perf_stat); + + stats = kzalloc(size, GFP_KERNEL); + if (!stats) + return -ENOMEM; + + stat = &stats->scm_statistic[0]; + memcpy(&stat->stat_id, "MemLife ", sizeof(stat->stat_id)); + stat->stat_val = 0; + + /* Fetch the fuel gauge and populate it in payload */ + rc = drc_pmem_query_stats(p, stats, 1); + if (rc < 0) { + dev_dbg(&p->pdev->dev, "Err(%d) fetching fuel gauge\n", rc); + goto free_stats; + } + + statval = be64_to_cpu(stat->stat_val); + dev_dbg(&p->pdev->dev, + "Fetched fuel-gauge %llu", statval); + payload->health.extension_flags |= + PDSM_DIMM_HEALTH_RUN_GAUGE_VALID; + payload->health.dimm_fuel_gauge = statval; + + rc = sizeof(struct nd_papr_pdsm_health); + +free_stats: + kfree(stats); + return rc; +} + +/* Add the dirty-shutdown-counter value to the pdsm */ +static int papr_pdsm_dsc(struct papr_scm_priv *p, + union nd_pdsm_payload *payload) +{ + payload->health.extension_flags |= PDSM_DIMM_DSC_VALID; + payload->health.dimm_dsc = p->dirty_shutdown_counter; + + return sizeof(struct nd_papr_pdsm_health); +} + +/* Fetch the DIMM health info and populate it in provided package. */ +static int papr_pdsm_health(struct papr_scm_priv *p, + union nd_pdsm_payload *payload) +{ + int rc; + + /* Ensure dimm health mutex is taken preventing concurrent access */ + rc = mutex_lock_interruptible(&p->health_mutex); + if (rc) + goto out; + + /* Always fetch upto date dimm health data ignoring cached values */ + rc = __drc_pmem_query_health(p); + if (rc) { + mutex_unlock(&p->health_mutex); + goto out; + } + + /* update health struct with various flags derived from health bitmap */ + payload->health = (struct nd_papr_pdsm_health) { + .extension_flags = 0, + .dimm_unarmed = !!(p->health_bitmap & PAPR_PMEM_UNARMED_MASK), + .dimm_bad_shutdown = !!(p->health_bitmap & PAPR_PMEM_BAD_SHUTDOWN_MASK), + .dimm_bad_restore = !!(p->health_bitmap & PAPR_PMEM_BAD_RESTORE_MASK), + .dimm_scrubbed = !!(p->health_bitmap & PAPR_PMEM_SCRUBBED_AND_LOCKED), + .dimm_locked = !!(p->health_bitmap & PAPR_PMEM_SCRUBBED_AND_LOCKED), + .dimm_encrypted = !!(p->health_bitmap & PAPR_PMEM_ENCRYPTED), + .dimm_health = PAPR_PDSM_DIMM_HEALTHY, + }; + + /* Update field dimm_health based on health_bitmap flags */ + if (p->health_bitmap & PAPR_PMEM_HEALTH_FATAL) + payload->health.dimm_health = PAPR_PDSM_DIMM_FATAL; + else if (p->health_bitmap & PAPR_PMEM_HEALTH_CRITICAL) + payload->health.dimm_health = PAPR_PDSM_DIMM_CRITICAL; + else if (p->health_bitmap & PAPR_PMEM_HEALTH_UNHEALTHY) + payload->health.dimm_health = PAPR_PDSM_DIMM_UNHEALTHY; + + /* struct populated hence can release the mutex now */ + mutex_unlock(&p->health_mutex); + + /* Populate the fuel gauge meter in the payload */ + papr_pdsm_fuel_gauge(p, payload); + /* Populate the dirty-shutdown-counter field */ + papr_pdsm_dsc(p, payload); + + rc = sizeof(struct nd_papr_pdsm_health); + +out: + return rc; +} + +/* Inject a smart error Add the dirty-shutdown-counter value to the pdsm */ +static int papr_pdsm_smart_inject(struct papr_scm_priv *p, + union nd_pdsm_payload *payload) +{ + int rc; + u32 supported_flags = 0; + u64 inject_mask = 0, clear_mask = 0; + u64 mask; + + /* Check for individual smart error flags and update inject/clear masks */ + if (payload->smart_inject.flags & PDSM_SMART_INJECT_HEALTH_FATAL) { + supported_flags |= PDSM_SMART_INJECT_HEALTH_FATAL; + if (payload->smart_inject.fatal_enable) + inject_mask |= PAPR_PMEM_HEALTH_FATAL; + else + clear_mask |= PAPR_PMEM_HEALTH_FATAL; + } + + if (payload->smart_inject.flags & PDSM_SMART_INJECT_BAD_SHUTDOWN) { + supported_flags |= PDSM_SMART_INJECT_BAD_SHUTDOWN; + if (payload->smart_inject.unsafe_shutdown_enable) + inject_mask |= PAPR_PMEM_SHUTDOWN_DIRTY; + else + clear_mask |= PAPR_PMEM_SHUTDOWN_DIRTY; + } + + dev_dbg(&p->pdev->dev, "[Smart-inject] inject_mask=%#llx clear_mask=%#llx\n", + inject_mask, clear_mask); + + /* Prevent concurrent access to dimm health bitmap related members */ + rc = mutex_lock_interruptible(&p->health_mutex); + if (rc) + return rc; + + /* Use inject/clear masks to set health_bitmap_inject_mask */ + mask = READ_ONCE(p->health_bitmap_inject_mask); + mask = (mask & ~clear_mask) | inject_mask; + WRITE_ONCE(p->health_bitmap_inject_mask, mask); + + /* Invalidate cached health bitmap */ + p->lasthealth_jiffies = 0; + + mutex_unlock(&p->health_mutex); + + /* Return the supported flags back to userspace */ + payload->smart_inject.flags = supported_flags; + + return sizeof(struct nd_papr_pdsm_health); +} + +/* + * 'struct pdsm_cmd_desc' + * Identifies supported PDSMs' expected length of in/out payloads + * and pdsm service function. + * + * size_in : Size of input payload if any in the PDSM request. + * size_out : Size of output payload if any in the PDSM request. + * service : Service function for the PDSM request. Return semantics: + * rc < 0 : Error servicing PDSM and rc indicates the error. + * rc >=0 : Serviced successfully and 'rc' indicate number of + * bytes written to payload. + */ +struct pdsm_cmd_desc { + u32 size_in; + u32 size_out; + int (*service)(struct papr_scm_priv *dimm, + union nd_pdsm_payload *payload); +}; + +/* Holds all supported PDSMs' command descriptors */ +static const struct pdsm_cmd_desc __pdsm_cmd_descriptors[] = { + [PAPR_PDSM_MIN] = { + .size_in = 0, + .size_out = 0, + .service = NULL, + }, + /* New PDSM command descriptors to be added below */ + + [PAPR_PDSM_HEALTH] = { + .size_in = 0, + .size_out = sizeof(struct nd_papr_pdsm_health), + .service = papr_pdsm_health, + }, + + [PAPR_PDSM_SMART_INJECT] = { + .size_in = sizeof(struct nd_papr_pdsm_smart_inject), + .size_out = sizeof(struct nd_papr_pdsm_smart_inject), + .service = papr_pdsm_smart_inject, + }, + /* Empty */ + [PAPR_PDSM_MAX] = { + .size_in = 0, + .size_out = 0, + .service = NULL, + }, +}; + +/* Given a valid pdsm cmd return its command descriptor else return NULL */ +static inline const struct pdsm_cmd_desc *pdsm_cmd_desc(enum papr_pdsm cmd) +{ + if (cmd >= 0 || cmd < ARRAY_SIZE(__pdsm_cmd_descriptors)) + return &__pdsm_cmd_descriptors[cmd]; + + return NULL; +} + +/* + * For a given pdsm request call an appropriate service function. + * Returns errors if any while handling the pdsm command package. + */ +static int papr_scm_service_pdsm(struct papr_scm_priv *p, + struct nd_cmd_pkg *pkg) +{ + /* Get the PDSM header and PDSM command */ + struct nd_pkg_pdsm *pdsm_pkg = (struct nd_pkg_pdsm *)pkg->nd_payload; + enum papr_pdsm pdsm = (enum papr_pdsm)pkg->nd_command; + const struct pdsm_cmd_desc *pdsc; + int rc; + + /* Fetch corresponding pdsm descriptor for validation and servicing */ + pdsc = pdsm_cmd_desc(pdsm); + + /* Validate pdsm descriptor */ + /* Ensure that reserved fields are 0 */ + if (pdsm_pkg->reserved[0] || pdsm_pkg->reserved[1]) { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Invalid reserved field\n", + pdsm); + return -EINVAL; + } + + /* If pdsm expects some input, then ensure that the size_in matches */ + if (pdsc->size_in && + pkg->nd_size_in != (pdsc->size_in + ND_PDSM_HDR_SIZE)) { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Mismatched size_in=%d\n", + pdsm, pkg->nd_size_in); + return -EINVAL; + } + + /* If pdsm wants to return data, then ensure that size_out matches */ + if (pdsc->size_out && + pkg->nd_size_out != (pdsc->size_out + ND_PDSM_HDR_SIZE)) { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Mismatched size_out=%d\n", + pdsm, pkg->nd_size_out); + return -EINVAL; + } + + /* Service the pdsm */ + if (pdsc->service) { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Servicing..\n", pdsm); + + rc = pdsc->service(p, &pdsm_pkg->payload); + + if (rc < 0) { + /* error encountered while servicing pdsm */ + pdsm_pkg->cmd_status = rc; + pkg->nd_fw_size = ND_PDSM_HDR_SIZE; + } else { + /* pdsm serviced and 'rc' bytes written to payload */ + pdsm_pkg->cmd_status = 0; + pkg->nd_fw_size = ND_PDSM_HDR_SIZE + rc; + } + } else { + dev_dbg(&p->pdev->dev, "PDSM[0x%x]: Unsupported PDSM request\n", + pdsm); + pdsm_pkg->cmd_status = -ENOENT; + pkg->nd_fw_size = ND_PDSM_HDR_SIZE; + } + + return pdsm_pkg->cmd_status; +} + +static int papr_scm_ndctl(struct nvdimm_bus_descriptor *nd_desc, + struct nvdimm *nvdimm, unsigned int cmd, void *buf, + unsigned int buf_len, int *cmd_rc) +{ + struct nd_cmd_get_config_size *get_size_hdr; + struct nd_cmd_pkg *call_pkg = NULL; + struct papr_scm_priv *p; + int rc; + + rc = is_cmd_valid(nvdimm, cmd, buf, buf_len); + if (rc) { + pr_debug("Invalid cmd=0x%x. Err=%d\n", cmd, rc); + return rc; + } + + /* Use a local variable in case cmd_rc pointer is NULL */ + if (!cmd_rc) + cmd_rc = &rc; + + p = nvdimm_provider_data(nvdimm); + + switch (cmd) { + case ND_CMD_GET_CONFIG_SIZE: + get_size_hdr = buf; + + get_size_hdr->status = 0; + get_size_hdr->max_xfer = 8; + get_size_hdr->config_size = p->metadata_size; + *cmd_rc = 0; + break; + + case ND_CMD_GET_CONFIG_DATA: + *cmd_rc = papr_scm_meta_get(p, buf); + break; + + case ND_CMD_SET_CONFIG_DATA: + *cmd_rc = papr_scm_meta_set(p, buf); + break; + + case ND_CMD_CALL: + call_pkg = (struct nd_cmd_pkg *)buf; + *cmd_rc = papr_scm_service_pdsm(p, call_pkg); + break; + + default: + dev_dbg(&p->pdev->dev, "Unknown command = %d\n", cmd); + return -EINVAL; + } + + dev_dbg(&p->pdev->dev, "returned with cmd_rc = %d\n", *cmd_rc); + + return 0; +} + +static ssize_t health_bitmap_inject_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct nvdimm *dimm = to_nvdimm(dev); + struct papr_scm_priv *p = nvdimm_provider_data(dimm); + + return sprintf(buf, "%#llx\n", + READ_ONCE(p->health_bitmap_inject_mask)); +} + +static DEVICE_ATTR_ADMIN_RO(health_bitmap_inject); + +static ssize_t perf_stats_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int index; + ssize_t rc; + struct seq_buf s; + struct papr_scm_perf_stat *stat; + struct papr_scm_perf_stats *stats; + struct nvdimm *dimm = to_nvdimm(dev); + struct papr_scm_priv *p = nvdimm_provider_data(dimm); + + if (!p->stat_buffer_len) + return -ENOENT; + + /* Allocate the buffer for phyp where stats are written */ + stats = kzalloc(p->stat_buffer_len, GFP_KERNEL); + if (!stats) + return -ENOMEM; + + /* Ask phyp to return all dimm perf stats */ + rc = drc_pmem_query_stats(p, stats, 0); + if (rc) + goto free_stats; + /* + * Go through the returned output buffer and print stats and + * values. Since stat_id is essentially a char string of + * 8 bytes, simply use the string format specifier to print it. + */ + seq_buf_init(&s, buf, PAGE_SIZE); + for (index = 0, stat = stats->scm_statistic; + index < be32_to_cpu(stats->num_statistics); + ++index, ++stat) { + seq_buf_printf(&s, "%.8s = 0x%016llX\n", + stat->stat_id, + be64_to_cpu(stat->stat_val)); + } + +free_stats: + kfree(stats); + return rc ? rc : (ssize_t)seq_buf_used(&s); +} +static DEVICE_ATTR_ADMIN_RO(perf_stats); + +static ssize_t flags_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct nvdimm *dimm = to_nvdimm(dev); + struct papr_scm_priv *p = nvdimm_provider_data(dimm); + struct seq_buf s; + u64 health; + int rc; + + rc = drc_pmem_query_health(p); + if (rc) + return rc; + + /* Copy health_bitmap locally, check masks & update out buffer */ + health = READ_ONCE(p->health_bitmap); + + seq_buf_init(&s, buf, PAGE_SIZE); + if (health & PAPR_PMEM_UNARMED_MASK) + seq_buf_printf(&s, "not_armed "); + + if (health & PAPR_PMEM_BAD_SHUTDOWN_MASK) + seq_buf_printf(&s, "flush_fail "); + + if (health & PAPR_PMEM_BAD_RESTORE_MASK) + seq_buf_printf(&s, "restore_fail "); + + if (health & PAPR_PMEM_ENCRYPTED) + seq_buf_printf(&s, "encrypted "); + + if (health & PAPR_PMEM_SMART_EVENT_MASK) + seq_buf_printf(&s, "smart_notify "); + + if (health & PAPR_PMEM_SCRUBBED_AND_LOCKED) + seq_buf_printf(&s, "scrubbed locked "); + + if (seq_buf_used(&s)) + seq_buf_printf(&s, "\n"); + + return seq_buf_used(&s); +} +DEVICE_ATTR_RO(flags); + +static ssize_t dirty_shutdown_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct nvdimm *dimm = to_nvdimm(dev); + struct papr_scm_priv *p = nvdimm_provider_data(dimm); + + return sysfs_emit(buf, "%llu\n", p->dirty_shutdown_counter); +} +DEVICE_ATTR_RO(dirty_shutdown); + +static umode_t papr_nd_attribute_visible(struct kobject *kobj, + struct attribute *attr, int n) +{ + struct device *dev = kobj_to_dev(kobj); + struct nvdimm *nvdimm = to_nvdimm(dev); + struct papr_scm_priv *p = nvdimm_provider_data(nvdimm); + + /* For if perf-stats not available remove perf_stats sysfs */ + if (attr == &dev_attr_perf_stats.attr && p->stat_buffer_len == 0) + return 0; + + return attr->mode; +} + +/* papr_scm specific dimm attributes */ +static struct attribute *papr_nd_attributes[] = { + &dev_attr_flags.attr, + &dev_attr_perf_stats.attr, + &dev_attr_dirty_shutdown.attr, + &dev_attr_health_bitmap_inject.attr, + NULL, +}; + +static const struct attribute_group papr_nd_attribute_group = { + .name = "papr", + .is_visible = papr_nd_attribute_visible, + .attrs = papr_nd_attributes, +}; + +static const struct attribute_group *papr_nd_attr_groups[] = { + &papr_nd_attribute_group, + NULL, +}; + +static int papr_scm_nvdimm_init(struct papr_scm_priv *p) +{ + struct device *dev = &p->pdev->dev; + struct nd_mapping_desc mapping; + struct nd_region_desc ndr_desc; + unsigned long dimm_flags; + int target_nid, online_nid; + + p->bus_desc.ndctl = papr_scm_ndctl; + p->bus_desc.module = THIS_MODULE; + p->bus_desc.of_node = p->pdev->dev.of_node; + p->bus_desc.provider_name = kstrdup(p->pdev->name, GFP_KERNEL); + + /* Set the dimm command family mask to accept PDSMs */ + set_bit(NVDIMM_FAMILY_PAPR, &p->bus_desc.dimm_family_mask); + + if (!p->bus_desc.provider_name) + return -ENOMEM; + + p->bus = nvdimm_bus_register(NULL, &p->bus_desc); + if (!p->bus) { + dev_err(dev, "Error creating nvdimm bus %pOF\n", p->dn); + kfree(p->bus_desc.provider_name); + return -ENXIO; + } + + dimm_flags = 0; + set_bit(NDD_LABELING, &dimm_flags); + + /* + * Check if the nvdimm is unarmed. No locking needed as we are still + * initializing. Ignore error encountered if any. + */ + __drc_pmem_query_health(p); + + if (p->health_bitmap & PAPR_PMEM_UNARMED_MASK) + set_bit(NDD_UNARMED, &dimm_flags); + + p->nvdimm = nvdimm_create(p->bus, p, papr_nd_attr_groups, + dimm_flags, PAPR_SCM_DIMM_CMD_MASK, 0, NULL); + if (!p->nvdimm) { + dev_err(dev, "Error creating DIMM object for %pOF\n", p->dn); + goto err; + } + + if (nvdimm_bus_check_dimm_count(p->bus, 1)) + goto err; + + /* now add the region */ + + memset(&mapping, 0, sizeof(mapping)); + mapping.nvdimm = p->nvdimm; + mapping.start = 0; + mapping.size = p->blocks * p->block_size; // XXX: potential overflow? + + memset(&ndr_desc, 0, sizeof(ndr_desc)); + target_nid = dev_to_node(&p->pdev->dev); + online_nid = numa_map_to_online_node(target_nid); + ndr_desc.numa_node = online_nid; + ndr_desc.target_node = target_nid; + ndr_desc.res = &p->res; + ndr_desc.of_node = p->dn; + ndr_desc.provider_data = p; + ndr_desc.mapping = &mapping; + ndr_desc.num_mappings = 1; + ndr_desc.nd_set = &p->nd_set; + + if (p->hcall_flush_required) { + set_bit(ND_REGION_ASYNC, &ndr_desc.flags); + ndr_desc.flush = papr_scm_pmem_flush; + } + + if (p->is_volatile) + p->region = nvdimm_volatile_region_create(p->bus, &ndr_desc); + else { + set_bit(ND_REGION_PERSIST_MEMCTRL, &ndr_desc.flags); + p->region = nvdimm_pmem_region_create(p->bus, &ndr_desc); + } + if (!p->region) { + dev_err(dev, "Error registering region %pR from %pOF\n", + ndr_desc.res, p->dn); + goto err; + } + if (target_nid != online_nid) + dev_info(dev, "Region registered with target node %d and online node %d", + target_nid, online_nid); + + mutex_lock(&papr_ndr_lock); + list_add_tail(&p->region_list, &papr_nd_regions); + mutex_unlock(&papr_ndr_lock); + + return 0; + +err: nvdimm_bus_unregister(p->bus); + kfree(p->bus_desc.provider_name); + return -ENXIO; +} + +static void papr_scm_add_badblock(struct nd_region *region, + struct nvdimm_bus *bus, u64 phys_addr) +{ + u64 aligned_addr = ALIGN_DOWN(phys_addr, L1_CACHE_BYTES); + + if (nvdimm_bus_add_badrange(bus, aligned_addr, L1_CACHE_BYTES)) { + pr_err("Bad block registration for 0x%llx failed\n", phys_addr); + return; + } + + pr_debug("Add memory range (0x%llx - 0x%llx) as bad range\n", + aligned_addr, aligned_addr + L1_CACHE_BYTES); + + nvdimm_region_notify(region, NVDIMM_REVALIDATE_POISON); +} + +static int handle_mce_ue(struct notifier_block *nb, unsigned long val, + void *data) +{ + struct machine_check_event *evt = data; + struct papr_scm_priv *p; + u64 phys_addr; + bool found = false; + + if (evt->error_type != MCE_ERROR_TYPE_UE) + return NOTIFY_DONE; + + if (list_empty(&papr_nd_regions)) + return NOTIFY_DONE; + + /* + * The physical address obtained here is PAGE_SIZE aligned, so get the + * exact address from the effective address + */ + phys_addr = evt->u.ue_error.physical_address + + (evt->u.ue_error.effective_address & ~PAGE_MASK); + + if (!evt->u.ue_error.physical_address_provided || + !is_zone_device_page(pfn_to_page(phys_addr >> PAGE_SHIFT))) + return NOTIFY_DONE; + + /* mce notifier is called from a process context, so mutex is safe */ + mutex_lock(&papr_ndr_lock); + list_for_each_entry(p, &papr_nd_regions, region_list) { + if (phys_addr >= p->res.start && phys_addr <= p->res.end) { + found = true; + break; + } + } + + if (found) + papr_scm_add_badblock(p->region, p->bus, phys_addr); + + mutex_unlock(&papr_ndr_lock); + + return found ? NOTIFY_OK : NOTIFY_DONE; +} + +static struct notifier_block mce_ue_nb = { + .notifier_call = handle_mce_ue +}; + +static int papr_scm_probe(struct platform_device *pdev) +{ + struct device_node *dn = pdev->dev.of_node; + u32 drc_index, metadata_size; + u64 blocks, block_size; + struct papr_scm_priv *p; + u8 uuid_raw[UUID_SIZE]; + const char *uuid_str; + ssize_t stat_size; + uuid_t uuid; + int rc; + + /* check we have all the required DT properties */ + if (of_property_read_u32(dn, "ibm,my-drc-index", &drc_index)) { + dev_err(&pdev->dev, "%pOF: missing drc-index!\n", dn); + return -ENODEV; + } + + if (of_property_read_u64(dn, "ibm,block-size", &block_size)) { + dev_err(&pdev->dev, "%pOF: missing block-size!\n", dn); + return -ENODEV; + } + + if (of_property_read_u64(dn, "ibm,number-of-blocks", &blocks)) { + dev_err(&pdev->dev, "%pOF: missing number-of-blocks!\n", dn); + return -ENODEV; + } + + if (of_property_read_string(dn, "ibm,unit-guid", &uuid_str)) { + dev_err(&pdev->dev, "%pOF: missing unit-guid!\n", dn); + return -ENODEV; + } + + /* + * open firmware platform device create won't update the NUMA + * distance table. For PAPR SCM devices we use numa_map_to_online_node() + * to find the nearest online NUMA node and that requires correct + * distance table information. + */ + update_numa_distance(dn); + + p = kzalloc(sizeof(*p), GFP_KERNEL); + if (!p) + return -ENOMEM; + + /* Initialize the dimm mutex */ + mutex_init(&p->health_mutex); + + /* optional DT properties */ + of_property_read_u32(dn, "ibm,metadata-size", &metadata_size); + + p->dn = dn; + p->drc_index = drc_index; + p->block_size = block_size; + p->blocks = blocks; + p->is_volatile = !of_property_read_bool(dn, "ibm,cache-flush-required"); + p->hcall_flush_required = of_property_read_bool(dn, "ibm,hcall-flush-required"); + + if (of_property_read_u64(dn, "ibm,persistence-failed-count", + &p->dirty_shutdown_counter)) + p->dirty_shutdown_counter = 0; + + /* We just need to ensure that set cookies are unique across */ + uuid_parse(uuid_str, &uuid); + + /* + * The cookie1 and cookie2 are not really little endian. + * We store a raw buffer representation of the + * uuid string so that we can compare this with the label + * area cookie irrespective of the endian configuration + * with which the kernel is built. + * + * Historically we stored the cookie in the below format. + * for a uuid string 72511b67-0b3b-42fd-8d1d-5be3cae8bcaa + * cookie1 was 0xfd423b0b671b5172 + * cookie2 was 0xaabce8cae35b1d8d + */ + export_uuid(uuid_raw, &uuid); + p->nd_set.cookie1 = get_unaligned_le64(&uuid_raw[0]); + p->nd_set.cookie2 = get_unaligned_le64(&uuid_raw[8]); + + /* might be zero */ + p->metadata_size = metadata_size; + p->pdev = pdev; + + /* request the hypervisor to bind this region to somewhere in memory */ + rc = drc_pmem_bind(p); + + /* If phyp says drc memory still bound then force unbound and retry */ + if (rc == H_OVERLAP) + rc = drc_pmem_query_n_bind(p); + + if (rc != H_SUCCESS) { + dev_err(&p->pdev->dev, "bind err: %d\n", rc); + rc = -ENXIO; + goto err; + } + + /* setup the resource for the newly bound range */ + p->res.start = p->bound_addr; + p->res.end = p->bound_addr + p->blocks * p->block_size - 1; + p->res.name = pdev->name; + p->res.flags = IORESOURCE_MEM; + + /* Try retrieving the stat buffer and see if its supported */ + stat_size = drc_pmem_query_stats(p, NULL, 0); + if (stat_size > 0) { + p->stat_buffer_len = stat_size; + dev_dbg(&p->pdev->dev, "Max perf-stat size %lu-bytes\n", + p->stat_buffer_len); + } + + rc = papr_scm_nvdimm_init(p); + if (rc) + goto err2; + + platform_set_drvdata(pdev, p); + papr_scm_pmu_register(p); + + return 0; + +err2: drc_pmem_unbind(p); +err: kfree(p); + return rc; +} + +static int papr_scm_remove(struct platform_device *pdev) +{ + struct papr_scm_priv *p = platform_get_drvdata(pdev); + + mutex_lock(&papr_ndr_lock); + list_del(&p->region_list); + mutex_unlock(&papr_ndr_lock); + + nvdimm_bus_unregister(p->bus); + drc_pmem_unbind(p); + + if (pdev->archdata.priv) + unregister_nvdimm_pmu(pdev->archdata.priv); + + pdev->archdata.priv = NULL; + kfree(p->bus_desc.provider_name); + kfree(p); + + return 0; +} + +static const struct of_device_id papr_scm_match[] = { + { .compatible = "ibm,pmemory" }, + { .compatible = "ibm,pmemory-v2" }, + { }, +}; + +static struct platform_driver papr_scm_driver = { + .probe = papr_scm_probe, + .remove = papr_scm_remove, + .driver = { + .name = "papr_scm", + .of_match_table = papr_scm_match, + }, +}; + +static int __init papr_scm_init(void) +{ + int ret; + + ret = platform_driver_register(&papr_scm_driver); + if (!ret) + mce_register_notifier(&mce_ue_nb); + + return ret; +} +module_init(papr_scm_init); + +static void __exit papr_scm_exit(void) +{ + mce_unregister_notifier(&mce_ue_nb); + platform_driver_unregister(&papr_scm_driver); +} +module_exit(papr_scm_exit); + +MODULE_DEVICE_TABLE(of, papr_scm_match); +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("IBM Corporation"); diff --git a/arch/powerpc/platforms/pseries/pci.c b/arch/powerpc/platforms/pseries/pci.c new file mode 100644 index 000000000..1772ae3d1 --- /dev/null +++ b/arch/powerpc/platforms/pseries/pci.c @@ -0,0 +1,322 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2001 Dave Engebretsen, IBM Corporation + * Copyright (C) 2003 Anton Blanchard <anton@au.ibm.com>, IBM + * + * pSeries specific routines for PCI. + */ + +#include <linux/init.h> +#include <linux/ioport.h> +#include <linux/kernel.h> +#include <linux/pci.h> +#include <linux/string.h> + +#include <asm/eeh.h> +#include <asm/pci-bridge.h> +#include <asm/ppc-pci.h> +#include <asm/pci.h> +#include "pseries.h" + +#if 0 +void pcibios_name_device(struct pci_dev *dev) +{ + struct device_node *dn; + + /* + * Add IBM loc code (slot) as a prefix to the device names for service + */ + dn = pci_device_to_OF_node(dev); + if (dn) { + const char *loc_code = of_get_property(dn, "ibm,loc-code", + NULL); + if (loc_code) { + int loc_len = strlen(loc_code); + if (loc_len < sizeof(dev->dev.name)) { + memmove(dev->dev.name+loc_len+1, dev->dev.name, + sizeof(dev->dev.name)-loc_len-1); + memcpy(dev->dev.name, loc_code, loc_len); + dev->dev.name[loc_len] = ' '; + dev->dev.name[sizeof(dev->dev.name)-1] = '\0'; + } + } + } +} +DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, pcibios_name_device); +#endif + +#ifdef CONFIG_PCI_IOV +#define MAX_VFS_FOR_MAP_PE 256 +struct pe_map_bar_entry { + __be64 bar; /* Input: Virtual Function BAR */ + __be16 rid; /* Input: Virtual Function Router ID */ + __be16 pe_num; /* Output: Virtual Function PE Number */ + __be32 reserved; /* Reserved Space */ +}; + +static int pseries_send_map_pe(struct pci_dev *pdev, u16 num_vfs, + struct pe_map_bar_entry *vf_pe_array) +{ + struct pci_dn *pdn; + int rc; + unsigned long buid, addr; + int ibm_map_pes = rtas_function_token(RTAS_FN_IBM_OPEN_SRIOV_MAP_PE_NUMBER); + + if (ibm_map_pes == RTAS_UNKNOWN_SERVICE) + return -EINVAL; + + pdn = pci_get_pdn(pdev); + addr = rtas_config_addr(pdn->busno, pdn->devfn, 0); + buid = pdn->phb->buid; + spin_lock(&rtas_data_buf_lock); + memcpy(rtas_data_buf, vf_pe_array, + RTAS_DATA_BUF_SIZE); + rc = rtas_call(ibm_map_pes, 5, 1, NULL, addr, + BUID_HI(buid), BUID_LO(buid), + rtas_data_buf, + num_vfs * sizeof(struct pe_map_bar_entry)); + memcpy(vf_pe_array, rtas_data_buf, RTAS_DATA_BUF_SIZE); + spin_unlock(&rtas_data_buf_lock); + + if (rc) + dev_err(&pdev->dev, + "%s: Failed to associate pes PE#%lx, rc=%x\n", + __func__, addr, rc); + + return rc; +} + +static void pseries_set_pe_num(struct pci_dev *pdev, u16 vf_index, __be16 pe_num) +{ + struct pci_dn *pdn; + + pdn = pci_get_pdn(pdev); + pdn->pe_num_map[vf_index] = be16_to_cpu(pe_num); + dev_dbg(&pdev->dev, "VF %04x:%02x:%02x.%x associated with PE#%x\n", + pci_domain_nr(pdev->bus), + pdev->bus->number, + PCI_SLOT(pci_iov_virtfn_devfn(pdev, vf_index)), + PCI_FUNC(pci_iov_virtfn_devfn(pdev, vf_index)), + pdn->pe_num_map[vf_index]); +} + +static int pseries_associate_pes(struct pci_dev *pdev, u16 num_vfs) +{ + struct pci_dn *pdn; + int i, rc, vf_index; + struct pe_map_bar_entry *vf_pe_array; + struct resource *res; + u64 size; + + vf_pe_array = kzalloc(RTAS_DATA_BUF_SIZE, GFP_KERNEL); + if (!vf_pe_array) + return -ENOMEM; + + pdn = pci_get_pdn(pdev); + /* create firmware structure to associate pes */ + for (vf_index = 0; vf_index < num_vfs; vf_index++) { + pdn->pe_num_map[vf_index] = IODA_INVALID_PE; + for (i = 0; i < PCI_SRIOV_NUM_BARS; i++) { + res = &pdev->resource[i + PCI_IOV_RESOURCES]; + if (!res->parent) + continue; + size = pcibios_iov_resource_alignment(pdev, i + + PCI_IOV_RESOURCES); + vf_pe_array[vf_index].bar = + cpu_to_be64(res->start + size * vf_index); + vf_pe_array[vf_index].rid = + cpu_to_be16((pci_iov_virtfn_bus(pdev, vf_index) + << 8) | pci_iov_virtfn_devfn(pdev, + vf_index)); + vf_pe_array[vf_index].pe_num = + cpu_to_be16(IODA_INVALID_PE); + } + } + + rc = pseries_send_map_pe(pdev, num_vfs, vf_pe_array); + /* Only zero is success */ + if (!rc) + for (vf_index = 0; vf_index < num_vfs; vf_index++) + pseries_set_pe_num(pdev, vf_index, + vf_pe_array[vf_index].pe_num); + + kfree(vf_pe_array); + return rc; +} + +static int pseries_pci_sriov_enable(struct pci_dev *pdev, u16 num_vfs) +{ + struct pci_dn *pdn; + int rc; + const int *max_vfs; + int max_config_vfs; + struct device_node *dn = pci_device_to_OF_node(pdev); + + max_vfs = of_get_property(dn, "ibm,number-of-configurable-vfs", NULL); + + if (!max_vfs) + return -EINVAL; + + /* First integer stores max config */ + max_config_vfs = of_read_number(&max_vfs[0], 1); + if (max_config_vfs < num_vfs && num_vfs > MAX_VFS_FOR_MAP_PE) { + dev_err(&pdev->dev, + "Num VFs %x > %x Configurable VFs\n", + num_vfs, (num_vfs > MAX_VFS_FOR_MAP_PE) ? + MAX_VFS_FOR_MAP_PE : max_config_vfs); + return -EINVAL; + } + + pdn = pci_get_pdn(pdev); + pdn->pe_num_map = kmalloc_array(num_vfs, + sizeof(*pdn->pe_num_map), + GFP_KERNEL); + if (!pdn->pe_num_map) + return -ENOMEM; + + rc = pseries_associate_pes(pdev, num_vfs); + + /* Anything other than zero is failure */ + if (rc) { + dev_err(&pdev->dev, "Failure to enable sriov: %x\n", rc); + kfree(pdn->pe_num_map); + } else { + pci_vf_drivers_autoprobe(pdev, false); + } + + return rc; +} + +static int pseries_pcibios_sriov_enable(struct pci_dev *pdev, u16 num_vfs) +{ + /* Allocate PCI data */ + add_sriov_vf_pdns(pdev); + return pseries_pci_sriov_enable(pdev, num_vfs); +} + +static int pseries_pcibios_sriov_disable(struct pci_dev *pdev) +{ + struct pci_dn *pdn; + + pdn = pci_get_pdn(pdev); + /* Releasing pe_num_map */ + kfree(pdn->pe_num_map); + /* Release PCI data */ + remove_sriov_vf_pdns(pdev); + pci_vf_drivers_autoprobe(pdev, true); + return 0; +} +#endif + +static void __init pSeries_request_regions(void) +{ + if (!isa_io_base) + return; + + request_region(0x20,0x20,"pic1"); + request_region(0xa0,0x20,"pic2"); + request_region(0x00,0x20,"dma1"); + request_region(0x40,0x20,"timer"); + request_region(0x80,0x10,"dma page reg"); + request_region(0xc0,0x20,"dma2"); +} + +void __init pSeries_final_fixup(void) +{ + pSeries_request_regions(); + + eeh_show_enabled(); + +#ifdef CONFIG_PCI_IOV + ppc_md.pcibios_sriov_enable = pseries_pcibios_sriov_enable; + ppc_md.pcibios_sriov_disable = pseries_pcibios_sriov_disable; +#endif +} + +/* + * Assume the winbond 82c105 is the IDE controller on a + * p610/p615/p630. We should probably be more careful in case + * someone tries to plug in a similar adapter. + */ +static void fixup_winbond_82c105(struct pci_dev* dev) +{ + struct resource *r; + unsigned int reg; + + if (!machine_is(pseries)) + return; + + printk("Using INTC for W82c105 IDE controller.\n"); + pci_read_config_dword(dev, 0x40, ®); + /* Enable LEGIRQ to use INTC instead of ISA interrupts */ + pci_write_config_dword(dev, 0x40, reg | (1<<11)); + + pci_dev_for_each_resource(dev, r) { + /* zap the 2nd function of the winbond chip */ + if (dev->bus->number == 0 && dev->devfn == 0x81 && + r->flags & IORESOURCE_IO) + r->flags &= ~IORESOURCE_IO; + if (r->start == 0 && r->end) { + r->flags = 0; + r->end = 0; + } + } +} +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_WINBOND, PCI_DEVICE_ID_WINBOND_82C105, + fixup_winbond_82c105); + +static enum pci_bus_speed prop_to_pci_speed(u32 prop) +{ + switch (prop) { + case 0x01: + return PCIE_SPEED_2_5GT; + case 0x02: + return PCIE_SPEED_5_0GT; + case 0x04: + return PCIE_SPEED_8_0GT; + case 0x08: + return PCIE_SPEED_16_0GT; + case 0x10: + return PCIE_SPEED_32_0GT; + default: + pr_debug("Unexpected PCI link speed property value\n"); + return PCI_SPEED_UNKNOWN; + } +} + +int pseries_root_bridge_prepare(struct pci_host_bridge *bridge) +{ + struct device_node *dn, *pdn; + struct pci_bus *bus; + u32 pcie_link_speed_stats[2]; + int rc; + + bus = bridge->bus; + + /* Rely on the pcibios_free_controller_deferred() callback. */ + pci_set_host_bridge_release(bridge, pcibios_free_controller_deferred, + (void *) pci_bus_to_host(bus)); + + dn = pcibios_get_phb_of_node(bus); + if (!dn) + return 0; + + for (pdn = dn; pdn != NULL; pdn = of_get_next_parent(pdn)) { + rc = of_property_read_u32_array(pdn, + "ibm,pcie-link-speed-stats", + &pcie_link_speed_stats[0], 2); + if (!rc) + break; + } + + of_node_put(pdn); + + if (rc) { + pr_debug("no ibm,pcie-link-speed-stats property\n"); + return 0; + } + + bus->max_bus_speed = prop_to_pci_speed(pcie_link_speed_stats[0]); + bus->cur_bus_speed = prop_to_pci_speed(pcie_link_speed_stats[1]); + return 0; +} diff --git a/arch/powerpc/platforms/pseries/pci_dlpar.c b/arch/powerpc/platforms/pseries/pci_dlpar.c new file mode 100644 index 000000000..4ba824568 --- /dev/null +++ b/arch/powerpc/platforms/pseries/pci_dlpar.c @@ -0,0 +1,111 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * PCI Dynamic LPAR, PCI Hot Plug and PCI EEH recovery code + * for RPA-compliant PPC64 platform. + * Copyright (C) 2003 Linda Xie <lxie@us.ibm.com> + * Copyright (C) 2005 International Business Machines + * + * Updates, 2005, John Rose <johnrose@austin.ibm.com> + * Updates, 2005, Linas Vepstas <linas@austin.ibm.com> + */ + +#include <linux/pci.h> +#include <linux/export.h> +#include <asm/pci-bridge.h> +#include <asm/ppc-pci.h> +#include <asm/firmware.h> +#include <asm/eeh.h> + +#include "pseries.h" + +struct pci_controller *init_phb_dynamic(struct device_node *dn) +{ + struct pci_controller *phb; + + pr_debug("PCI: Initializing new hotplug PHB %pOF\n", dn); + + phb = pcibios_alloc_controller(dn); + if (!phb) + return NULL; + rtas_setup_phb(phb); + pci_process_bridge_OF_ranges(phb, dn, 0); + phb->controller_ops = pseries_pci_controller_ops; + + pci_devs_phb_init_dynamic(phb); + + pseries_msi_allocate_domains(phb); + + /* Create EEH devices for the PHB */ + eeh_phb_pe_create(phb); + + if (dn->child) + pseries_eeh_init_edev_recursive(PCI_DN(dn)); + + pcibios_scan_phb(phb); + pcibios_finish_adding_to_bus(phb->bus); + + return phb; +} +EXPORT_SYMBOL_GPL(init_phb_dynamic); + +/* RPA-specific bits for removing PHBs */ +int remove_phb_dynamic(struct pci_controller *phb) +{ + struct pci_bus *b = phb->bus; + struct pci_host_bridge *host_bridge = to_pci_host_bridge(b->bridge); + struct resource *res; + int rc, i; + + pr_debug("PCI: Removing PHB %04x:%02x...\n", + pci_domain_nr(b), b->number); + + /* We cannot to remove a root bus that has children */ + if (!(list_empty(&b->children) && list_empty(&b->devices))) + return -EBUSY; + + /* We -know- there aren't any child devices anymore at this stage + * and thus, we can safely unmap the IO space as it's not in use + */ + res = &phb->io_resource; + if (res->flags & IORESOURCE_IO) { + rc = pcibios_unmap_io_space(b); + if (rc) { + printk(KERN_ERR "%s: failed to unmap IO on bus %s\n", + __func__, b->name); + return 1; + } + } + + pseries_msi_free_domains(phb); + + /* Keep a reference so phb isn't freed yet */ + get_device(&host_bridge->dev); + + /* Remove the PCI bus and unregister the bridge device from sysfs */ + phb->bus = NULL; + pci_remove_bus(b); + host_bridge->bus = NULL; + device_unregister(&host_bridge->dev); + + /* Now release the IO resource */ + if (res->flags & IORESOURCE_IO) + release_resource(res); + + /* Release memory resources */ + for (i = 0; i < 3; ++i) { + res = &phb->mem_resources[i]; + if (!(res->flags & IORESOURCE_MEM)) + continue; + release_resource(res); + } + + /* + * The pci_controller data structure is freed by + * the pcibios_free_controller_deferred() callback; + * see pseries_root_bridge_prepare(). + */ + put_device(&host_bridge->dev); + + return 0; +} +EXPORT_SYMBOL_GPL(remove_phb_dynamic); diff --git a/arch/powerpc/platforms/pseries/plpks-secvar.c b/arch/powerpc/platforms/pseries/plpks-secvar.c new file mode 100644 index 000000000..257fd1f8b --- /dev/null +++ b/arch/powerpc/platforms/pseries/plpks-secvar.c @@ -0,0 +1,217 @@ +// SPDX-License-Identifier: GPL-2.0-only + +// Secure variable implementation using the PowerVM LPAR Platform KeyStore (PLPKS) +// +// Copyright 2022, 2023 IBM Corporation +// Authors: Russell Currey +// Andrew Donnellan +// Nayna Jain + +#define pr_fmt(fmt) "secvar: "fmt + +#include <linux/printk.h> +#include <linux/init.h> +#include <linux/types.h> +#include <linux/slab.h> +#include <linux/string.h> +#include <linux/kobject.h> +#include <linux/nls.h> +#include <asm/machdep.h> +#include <asm/secvar.h> +#include <asm/plpks.h> + +// Config attributes for sysfs +#define PLPKS_CONFIG_ATTR(name, fmt, func) \ + static ssize_t name##_show(struct kobject *kobj, \ + struct kobj_attribute *attr, \ + char *buf) \ + { \ + return sysfs_emit(buf, fmt, func()); \ + } \ + static struct kobj_attribute attr_##name = __ATTR_RO(name) + +PLPKS_CONFIG_ATTR(version, "%u\n", plpks_get_version); +PLPKS_CONFIG_ATTR(max_object_size, "%u\n", plpks_get_maxobjectsize); +PLPKS_CONFIG_ATTR(total_size, "%u\n", plpks_get_totalsize); +PLPKS_CONFIG_ATTR(used_space, "%u\n", plpks_get_usedspace); +PLPKS_CONFIG_ATTR(supported_policies, "%08x\n", plpks_get_supportedpolicies); +PLPKS_CONFIG_ATTR(signed_update_algorithms, "%016llx\n", plpks_get_signedupdatealgorithms); + +static const struct attribute *config_attrs[] = { + &attr_version.attr, + &attr_max_object_size.attr, + &attr_total_size.attr, + &attr_used_space.attr, + &attr_supported_policies.attr, + &attr_signed_update_algorithms.attr, + NULL, +}; + +static u32 get_policy(const char *name) +{ + if ((strcmp(name, "db") == 0) || + (strcmp(name, "dbx") == 0) || + (strcmp(name, "grubdb") == 0) || + (strcmp(name, "grubdbx") == 0) || + (strcmp(name, "sbat") == 0)) + return (PLPKS_WORLDREADABLE | PLPKS_SIGNEDUPDATE); + else + return PLPKS_SIGNEDUPDATE; +} + +static const char * const plpks_var_names[] = { + "PK", + "KEK", + "db", + "dbx", + "grubdb", + "grubdbx", + "sbat", + "moduledb", + "trustedcadb", + NULL, +}; + +static int plpks_get_variable(const char *key, u64 key_len, u8 *data, + u64 *data_size) +{ + struct plpks_var var = {0}; + int rc = 0; + + // We subtract 1 from key_len because we don't need to include the + // null terminator at the end of the string + var.name = kcalloc(key_len - 1, sizeof(wchar_t), GFP_KERNEL); + if (!var.name) + return -ENOMEM; + rc = utf8s_to_utf16s(key, key_len - 1, UTF16_LITTLE_ENDIAN, (wchar_t *)var.name, + key_len - 1); + if (rc < 0) + goto err; + var.namelen = rc * 2; + + var.os = PLPKS_VAR_LINUX; + if (data) { + var.data = data; + var.datalen = *data_size; + } + rc = plpks_read_os_var(&var); + + if (rc) + goto err; + + *data_size = var.datalen; + +err: + kfree(var.name); + if (rc && rc != -ENOENT) { + pr_err("Failed to read variable '%s': %d\n", key, rc); + // Return -EIO since userspace probably doesn't care about the + // specific error + rc = -EIO; + } + return rc; +} + +static int plpks_set_variable(const char *key, u64 key_len, u8 *data, + u64 data_size) +{ + struct plpks_var var = {0}; + int rc = 0; + u64 flags; + + // Secure variables need to be prefixed with 8 bytes of flags. + // We only want to perform the write if we have at least one byte of data. + if (data_size <= sizeof(flags)) + return -EINVAL; + + // We subtract 1 from key_len because we don't need to include the + // null terminator at the end of the string + var.name = kcalloc(key_len - 1, sizeof(wchar_t), GFP_KERNEL); + if (!var.name) + return -ENOMEM; + rc = utf8s_to_utf16s(key, key_len - 1, UTF16_LITTLE_ENDIAN, (wchar_t *)var.name, + key_len - 1); + if (rc < 0) + goto err; + var.namelen = rc * 2; + + // Flags are contained in the first 8 bytes of the buffer, and are always big-endian + flags = be64_to_cpup((__be64 *)data); + + var.datalen = data_size - sizeof(flags); + var.data = data + sizeof(flags); + var.os = PLPKS_VAR_LINUX; + var.policy = get_policy(key); + + // Unlike in the read case, the plpks error code can be useful to + // userspace on write, so we return it rather than just -EIO + rc = plpks_signed_update_var(&var, flags); + +err: + kfree(var.name); + return rc; +} + +// PLPKS dynamic secure boot doesn't give us a format string in the same way OPAL does. +// Instead, report the format using the SB_VERSION variable in the keystore. +// The string is made up by us, and takes the form "ibm,plpks-sb-v<n>" (or "ibm,plpks-sb-unknown" +// if the SB_VERSION variable doesn't exist). Hypervisor defines the SB_VERSION variable as a +// "1 byte unsigned integer value". +static ssize_t plpks_secvar_format(char *buf, size_t bufsize) +{ + struct plpks_var var = {0}; + ssize_t ret; + u8 version; + + var.component = NULL; + // Only the signed variables have null bytes in their names, this one doesn't + var.name = "SB_VERSION"; + var.namelen = strlen(var.name); + var.datalen = 1; + var.data = &version; + + // Unlike the other vars, SB_VERSION is owned by firmware instead of the OS + ret = plpks_read_fw_var(&var); + if (ret) { + if (ret == -ENOENT) { + ret = snprintf(buf, bufsize, "ibm,plpks-sb-unknown"); + } else { + pr_err("Error %ld reading SB_VERSION from firmware\n", ret); + ret = -EIO; + } + goto err; + } + + ret = snprintf(buf, bufsize, "ibm,plpks-sb-v%hhu", version); +err: + return ret; +} + +static int plpks_max_size(u64 *max_size) +{ + // The max object size reported by the hypervisor is accurate for the + // object itself, but we use the first 8 bytes of data on write as the + // signed update flags, so the max size a user can write is larger. + *max_size = (u64)plpks_get_maxobjectsize() + sizeof(u64); + + return 0; +} + + +static const struct secvar_operations plpks_secvar_ops = { + .get = plpks_get_variable, + .set = plpks_set_variable, + .format = plpks_secvar_format, + .max_size = plpks_max_size, + .config_attrs = config_attrs, + .var_names = plpks_var_names, +}; + +static int plpks_secvar_init(void) +{ + if (!plpks_is_available()) + return -ENODEV; + + return set_secvar_ops(&plpks_secvar_ops); +} +machine_device_initcall(pseries, plpks_secvar_init); diff --git a/arch/powerpc/platforms/pseries/plpks.c b/arch/powerpc/platforms/pseries/plpks.c new file mode 100644 index 000000000..2d40304eb --- /dev/null +++ b/arch/powerpc/platforms/pseries/plpks.c @@ -0,0 +1,711 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * POWER LPAR Platform KeyStore(PLPKS) + * Copyright (C) 2022 IBM Corporation + * Author: Nayna Jain <nayna@linux.ibm.com> + * + * Provides access to variables stored in Power LPAR Platform KeyStore(PLPKS). + */ + +#define pr_fmt(fmt) "plpks: " fmt + +#include <linux/delay.h> +#include <linux/errno.h> +#include <linux/io.h> +#include <linux/printk.h> +#include <linux/slab.h> +#include <linux/string.h> +#include <linux/types.h> +#include <linux/of_fdt.h> +#include <linux/libfdt.h> +#include <linux/memblock.h> +#include <asm/hvcall.h> +#include <asm/machdep.h> +#include <asm/plpks.h> +#include <asm/firmware.h> + +static u8 *ospassword; +static u16 ospasswordlength; + +// Retrieved with H_PKS_GET_CONFIG +static u8 version; +static u16 objoverhead; +static u16 maxpwsize; +static u16 maxobjsize; +static s16 maxobjlabelsize; +static u32 totalsize; +static u32 usedspace; +static u32 supportedpolicies; +static u32 maxlargeobjectsize; +static u64 signedupdatealgorithms; + +struct plpks_auth { + u8 version; + u8 consumer; + __be64 rsvd0; + __be32 rsvd1; + __be16 passwordlength; + u8 password[]; +} __packed __aligned(16); + +struct label_attr { + u8 prefix[8]; + u8 version; + u8 os; + u8 length; + u8 reserved[5]; +}; + +struct label { + struct label_attr attr; + u8 name[PLPKS_MAX_NAME_SIZE]; + size_t size; +}; + +static int pseries_status_to_err(int rc) +{ + int err; + + switch (rc) { + case H_SUCCESS: + err = 0; + break; + case H_FUNCTION: + err = -ENXIO; + break; + case H_PARAMETER: + case H_P2: + case H_P3: + case H_P4: + case H_P5: + case H_P6: + err = -EINVAL; + break; + case H_NOT_FOUND: + err = -ENOENT; + break; + case H_BUSY: + case H_LONG_BUSY_ORDER_1_MSEC: + case H_LONG_BUSY_ORDER_10_MSEC: + case H_LONG_BUSY_ORDER_100_MSEC: + case H_LONG_BUSY_ORDER_1_SEC: + case H_LONG_BUSY_ORDER_10_SEC: + case H_LONG_BUSY_ORDER_100_SEC: + err = -EBUSY; + break; + case H_AUTHORITY: + err = -EPERM; + break; + case H_NO_MEM: + err = -ENOMEM; + break; + case H_RESOURCE: + err = -EEXIST; + break; + case H_TOO_BIG: + err = -EFBIG; + break; + case H_STATE: + err = -EIO; + break; + case H_R_STATE: + err = -EIO; + break; + case H_IN_USE: + err = -EEXIST; + break; + case H_ABORTED: + err = -EIO; + break; + default: + err = -EINVAL; + } + + pr_debug("Converted hypervisor code %d to Linux %d\n", rc, err); + + return err; +} + +static int plpks_gen_password(void) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE] = { 0 }; + u8 *password, consumer = PLPKS_OS_OWNER; + int rc; + + // If we booted from kexec, we could be reusing an existing password already + if (ospassword) { + pr_debug("Password of length %u already in use\n", ospasswordlength); + return 0; + } + + // The password must not cross a page boundary, so we align to the next power of 2 + password = kzalloc(roundup_pow_of_two(maxpwsize), GFP_KERNEL); + if (!password) + return -ENOMEM; + + rc = plpar_hcall(H_PKS_GEN_PASSWORD, retbuf, consumer, 0, + virt_to_phys(password), maxpwsize); + + if (!rc) { + ospasswordlength = maxpwsize; + ospassword = kzalloc(maxpwsize, GFP_KERNEL); + if (!ospassword) { + kfree(password); + return -ENOMEM; + } + memcpy(ospassword, password, ospasswordlength); + } else { + if (rc == H_IN_USE) { + pr_warn("Password already set - authenticated operations will fail\n"); + rc = 0; + } else { + goto out; + } + } +out: + kfree(password); + + return pseries_status_to_err(rc); +} + +static struct plpks_auth *construct_auth(u8 consumer) +{ + struct plpks_auth *auth; + + if (consumer > PLPKS_OS_OWNER) + return ERR_PTR(-EINVAL); + + // The auth structure must not cross a page boundary and must be + // 16 byte aligned. We align to the next largest power of 2 + auth = kzalloc(roundup_pow_of_two(struct_size(auth, password, maxpwsize)), GFP_KERNEL); + if (!auth) + return ERR_PTR(-ENOMEM); + + auth->version = 1; + auth->consumer = consumer; + + if (consumer == PLPKS_FW_OWNER || consumer == PLPKS_BOOTLOADER_OWNER) + return auth; + + memcpy(auth->password, ospassword, ospasswordlength); + + auth->passwordlength = cpu_to_be16(ospasswordlength); + + return auth; +} + +/* + * Label is combination of label attributes + name. + * Label attributes are used internally by kernel and not exposed to the user. + */ +static struct label *construct_label(char *component, u8 varos, u8 *name, + u16 namelen) +{ + struct label *label; + size_t slen = 0; + + if (!name || namelen > PLPKS_MAX_NAME_SIZE) + return ERR_PTR(-EINVAL); + + // Support NULL component for signed updates + if (component) { + slen = strlen(component); + if (slen > sizeof(label->attr.prefix)) + return ERR_PTR(-EINVAL); + } + + // The label structure must not cross a page boundary, so we align to the next power of 2 + label = kzalloc(roundup_pow_of_two(sizeof(*label)), GFP_KERNEL); + if (!label) + return ERR_PTR(-ENOMEM); + + if (component) + memcpy(&label->attr.prefix, component, slen); + + label->attr.version = PLPKS_LABEL_VERSION; + label->attr.os = varos; + label->attr.length = PLPKS_MAX_LABEL_ATTR_SIZE; + memcpy(&label->name, name, namelen); + + label->size = sizeof(struct label_attr) + namelen; + + return label; +} + +static int _plpks_get_config(void) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE] = { 0 }; + struct config { + u8 version; + u8 flags; + __be16 rsvd0; + __be16 objoverhead; + __be16 maxpwsize; + __be16 maxobjlabelsize; + __be16 maxobjsize; + __be32 totalsize; + __be32 usedspace; + __be32 supportedpolicies; + __be32 maxlargeobjectsize; + __be64 signedupdatealgorithms; + u8 rsvd1[476]; + } __packed * config; + size_t size; + int rc = 0; + + size = sizeof(*config); + + // Config struct must not cross a page boundary. So long as the struct + // size is a power of 2, this should be fine as alignment is guaranteed + config = kzalloc(size, GFP_KERNEL); + if (!config) { + rc = -ENOMEM; + goto err; + } + + rc = plpar_hcall(H_PKS_GET_CONFIG, retbuf, virt_to_phys(config), size); + + if (rc != H_SUCCESS) { + rc = pseries_status_to_err(rc); + goto err; + } + + version = config->version; + objoverhead = be16_to_cpu(config->objoverhead); + maxpwsize = be16_to_cpu(config->maxpwsize); + maxobjsize = be16_to_cpu(config->maxobjsize); + maxobjlabelsize = be16_to_cpu(config->maxobjlabelsize); + totalsize = be32_to_cpu(config->totalsize); + usedspace = be32_to_cpu(config->usedspace); + supportedpolicies = be32_to_cpu(config->supportedpolicies); + maxlargeobjectsize = be32_to_cpu(config->maxlargeobjectsize); + signedupdatealgorithms = be64_to_cpu(config->signedupdatealgorithms); + + // Validate that the numbers we get back match the requirements of the spec + if (maxpwsize < 32) { + pr_err("Invalid Max Password Size received from hypervisor (%d < 32)\n", maxpwsize); + rc = -EIO; + goto err; + } + + if (maxobjlabelsize < 255) { + pr_err("Invalid Max Object Label Size received from hypervisor (%d < 255)\n", + maxobjlabelsize); + rc = -EIO; + goto err; + } + + if (totalsize < 4096) { + pr_err("Invalid Total Size received from hypervisor (%d < 4096)\n", totalsize); + rc = -EIO; + goto err; + } + + if (version >= 3 && maxlargeobjectsize >= 65536 && maxobjsize != 0xFFFF) { + pr_err("Invalid Max Object Size (0x%x != 0xFFFF)\n", maxobjsize); + rc = -EIO; + goto err; + } + +err: + kfree(config); + return rc; +} + +u8 plpks_get_version(void) +{ + return version; +} + +u16 plpks_get_objoverhead(void) +{ + return objoverhead; +} + +u16 plpks_get_maxpwsize(void) +{ + return maxpwsize; +} + +u16 plpks_get_maxobjectsize(void) +{ + return maxobjsize; +} + +u16 plpks_get_maxobjectlabelsize(void) +{ + return maxobjlabelsize; +} + +u32 plpks_get_totalsize(void) +{ + return totalsize; +} + +u32 plpks_get_usedspace(void) +{ + // Unlike other config values, usedspace regularly changes as objects + // are updated, so we need to refresh. + int rc = _plpks_get_config(); + if (rc) { + pr_err("Couldn't get config, rc: %d\n", rc); + return 0; + } + return usedspace; +} + +u32 plpks_get_supportedpolicies(void) +{ + return supportedpolicies; +} + +u32 plpks_get_maxlargeobjectsize(void) +{ + return maxlargeobjectsize; +} + +u64 plpks_get_signedupdatealgorithms(void) +{ + return signedupdatealgorithms; +} + +u16 plpks_get_passwordlen(void) +{ + return ospasswordlength; +} + +bool plpks_is_available(void) +{ + int rc; + + if (!firmware_has_feature(FW_FEATURE_PLPKS)) + return false; + + rc = _plpks_get_config(); + if (rc) + return false; + + return true; +} + +static int plpks_confirm_object_flushed(struct label *label, + struct plpks_auth *auth) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE] = { 0 }; + bool timed_out = true; + u64 timeout = 0; + u8 status; + int rc; + + do { + rc = plpar_hcall(H_PKS_CONFIRM_OBJECT_FLUSHED, retbuf, + virt_to_phys(auth), virt_to_phys(label), + label->size); + + status = retbuf[0]; + if (rc) { + timed_out = false; + if (rc == H_NOT_FOUND && status == 1) + rc = 0; + break; + } + + if (!rc && status == 1) { + timed_out = false; + break; + } + + usleep_range(PLPKS_FLUSH_SLEEP, + PLPKS_FLUSH_SLEEP + PLPKS_FLUSH_SLEEP_RANGE); + timeout = timeout + PLPKS_FLUSH_SLEEP; + } while (timeout < PLPKS_MAX_TIMEOUT); + + if (timed_out) + return -ETIMEDOUT; + + return pseries_status_to_err(rc); +} + +int plpks_signed_update_var(struct plpks_var *var, u64 flags) +{ + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE] = {0}; + int rc; + struct label *label; + struct plpks_auth *auth; + u64 continuetoken = 0; + u64 timeout = 0; + + if (!var->data || var->datalen <= 0 || var->namelen > PLPKS_MAX_NAME_SIZE) + return -EINVAL; + + if (!(var->policy & PLPKS_SIGNEDUPDATE)) + return -EINVAL; + + // Signed updates need the component to be NULL. + if (var->component) + return -EINVAL; + + auth = construct_auth(PLPKS_OS_OWNER); + if (IS_ERR(auth)) + return PTR_ERR(auth); + + label = construct_label(var->component, var->os, var->name, var->namelen); + if (IS_ERR(label)) { + rc = PTR_ERR(label); + goto out; + } + + do { + rc = plpar_hcall9(H_PKS_SIGNED_UPDATE, retbuf, + virt_to_phys(auth), virt_to_phys(label), + label->size, var->policy, flags, + virt_to_phys(var->data), var->datalen, + continuetoken); + + continuetoken = retbuf[0]; + if (pseries_status_to_err(rc) == -EBUSY) { + int delay_ms = get_longbusy_msecs(rc); + mdelay(delay_ms); + timeout += delay_ms; + } + rc = pseries_status_to_err(rc); + } while (rc == -EBUSY && timeout < PLPKS_MAX_TIMEOUT); + + if (!rc) + rc = plpks_confirm_object_flushed(label, auth); + + kfree(label); +out: + kfree(auth); + + return rc; +} + +int plpks_write_var(struct plpks_var var) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE] = { 0 }; + struct plpks_auth *auth; + struct label *label; + int rc; + + if (!var.component || !var.data || var.datalen <= 0 || + var.namelen > PLPKS_MAX_NAME_SIZE || var.datalen > PLPKS_MAX_DATA_SIZE) + return -EINVAL; + + if (var.policy & PLPKS_SIGNEDUPDATE) + return -EINVAL; + + auth = construct_auth(PLPKS_OS_OWNER); + if (IS_ERR(auth)) + return PTR_ERR(auth); + + label = construct_label(var.component, var.os, var.name, var.namelen); + if (IS_ERR(label)) { + rc = PTR_ERR(label); + goto out; + } + + rc = plpar_hcall(H_PKS_WRITE_OBJECT, retbuf, virt_to_phys(auth), + virt_to_phys(label), label->size, var.policy, + virt_to_phys(var.data), var.datalen); + + if (!rc) + rc = plpks_confirm_object_flushed(label, auth); + + rc = pseries_status_to_err(rc); + kfree(label); +out: + kfree(auth); + + return rc; +} + +int plpks_remove_var(char *component, u8 varos, struct plpks_var_name vname) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE] = { 0 }; + struct plpks_auth *auth; + struct label *label; + int rc; + + if (vname.namelen > PLPKS_MAX_NAME_SIZE) + return -EINVAL; + + auth = construct_auth(PLPKS_OS_OWNER); + if (IS_ERR(auth)) + return PTR_ERR(auth); + + label = construct_label(component, varos, vname.name, vname.namelen); + if (IS_ERR(label)) { + rc = PTR_ERR(label); + goto out; + } + + rc = plpar_hcall(H_PKS_REMOVE_OBJECT, retbuf, virt_to_phys(auth), + virt_to_phys(label), label->size); + + if (!rc) + rc = plpks_confirm_object_flushed(label, auth); + + rc = pseries_status_to_err(rc); + kfree(label); +out: + kfree(auth); + + return rc; +} + +static int plpks_read_var(u8 consumer, struct plpks_var *var) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE] = { 0 }; + struct plpks_auth *auth; + struct label *label = NULL; + u8 *output; + int rc; + + if (var->namelen > PLPKS_MAX_NAME_SIZE) + return -EINVAL; + + auth = construct_auth(consumer); + if (IS_ERR(auth)) + return PTR_ERR(auth); + + if (consumer == PLPKS_OS_OWNER) { + label = construct_label(var->component, var->os, var->name, + var->namelen); + if (IS_ERR(label)) { + rc = PTR_ERR(label); + goto out_free_auth; + } + } + + output = kzalloc(maxobjsize, GFP_KERNEL); + if (!output) { + rc = -ENOMEM; + goto out_free_label; + } + + if (consumer == PLPKS_OS_OWNER) + rc = plpar_hcall(H_PKS_READ_OBJECT, retbuf, virt_to_phys(auth), + virt_to_phys(label), label->size, virt_to_phys(output), + maxobjsize); + else + rc = plpar_hcall(H_PKS_READ_OBJECT, retbuf, virt_to_phys(auth), + virt_to_phys(var->name), var->namelen, virt_to_phys(output), + maxobjsize); + + + if (rc != H_SUCCESS) { + rc = pseries_status_to_err(rc); + goto out_free_output; + } + + if (!var->data || var->datalen > retbuf[0]) + var->datalen = retbuf[0]; + + var->policy = retbuf[1]; + + if (var->data) + memcpy(var->data, output, var->datalen); + + rc = 0; + +out_free_output: + kfree(output); +out_free_label: + kfree(label); +out_free_auth: + kfree(auth); + + return rc; +} + +int plpks_read_os_var(struct plpks_var *var) +{ + return plpks_read_var(PLPKS_OS_OWNER, var); +} + +int plpks_read_fw_var(struct plpks_var *var) +{ + return plpks_read_var(PLPKS_FW_OWNER, var); +} + +int plpks_read_bootloader_var(struct plpks_var *var) +{ + return plpks_read_var(PLPKS_BOOTLOADER_OWNER, var); +} + +int plpks_populate_fdt(void *fdt) +{ + int chosen_offset = fdt_path_offset(fdt, "/chosen"); + + if (chosen_offset < 0) { + pr_err("Can't find chosen node: %s\n", + fdt_strerror(chosen_offset)); + return chosen_offset; + } + + return fdt_setprop(fdt, chosen_offset, "ibm,plpks-pw", ospassword, ospasswordlength); +} + +// Once a password is registered with the hypervisor it cannot be cleared without +// rebooting the LPAR, so to keep using the PLPKS across kexec boots we need to +// recover the previous password from the FDT. +// +// There are a few challenges here. We don't want the password to be visible to +// users, so we need to clear it from the FDT. This has to be done in early boot. +// Clearing it from the FDT would make the FDT's checksum invalid, so we have to +// manually cause the checksum to be recalculated. +void __init plpks_early_init_devtree(void) +{ + void *fdt = initial_boot_params; + int chosen_node = fdt_path_offset(fdt, "/chosen"); + const u8 *password; + int len; + + if (chosen_node < 0) + return; + + password = fdt_getprop(fdt, chosen_node, "ibm,plpks-pw", &len); + if (len <= 0) { + pr_debug("Couldn't find ibm,plpks-pw node.\n"); + return; + } + + ospassword = memblock_alloc_raw(len, SMP_CACHE_BYTES); + if (!ospassword) { + pr_err("Error allocating memory for password.\n"); + goto out; + } + + memcpy(ospassword, password, len); + ospasswordlength = (u16)len; + +out: + fdt_nop_property(fdt, chosen_node, "ibm,plpks-pw"); + // Since we've cleared the password, we must update the FDT checksum + early_init_dt_verify(fdt); +} + +static __init int pseries_plpks_init(void) +{ + int rc; + + if (!firmware_has_feature(FW_FEATURE_PLPKS)) + return -ENODEV; + + rc = _plpks_get_config(); + + if (rc) { + pr_err("POWER LPAR Platform KeyStore is not supported or enabled\n"); + return rc; + } + + rc = plpks_gen_password(); + if (rc) + pr_err("Failed setting POWER LPAR Platform KeyStore Password\n"); + else + pr_info("POWER LPAR Platform KeyStore initialized successfully\n"); + + return rc; +} +machine_arch_initcall(pseries, pseries_plpks_init); diff --git a/arch/powerpc/platforms/pseries/pmem.c b/arch/powerpc/platforms/pseries/pmem.c new file mode 100644 index 000000000..3c290b9ed --- /dev/null +++ b/arch/powerpc/platforms/pseries/pmem.c @@ -0,0 +1,167 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Handles hot and cold plug of persistent memory regions on pseries. + */ + +#define pr_fmt(fmt) "pseries-pmem: " fmt + +#include <linux/kernel.h> +#include <linux/interrupt.h> +#include <linux/delay.h> +#include <linux/sched.h> /* for idle_task_exit */ +#include <linux/sched/hotplug.h> +#include <linux/cpu.h> +#include <linux/of.h> +#include <linux/of_platform.h> +#include <linux/slab.h> +#include <asm/rtas.h> +#include <asm/firmware.h> +#include <asm/machdep.h> +#include <asm/vdso_datapage.h> +#include <asm/plpar_wrappers.h> +#include <asm/topology.h> + +#include "pseries.h" + +static struct device_node *pmem_node; + +static ssize_t pmem_drc_add_node(u32 drc_index) +{ + struct device_node *dn; + int rc; + + pr_debug("Attempting to add pmem node, drc index: %x\n", drc_index); + + rc = dlpar_acquire_drc(drc_index); + if (rc) { + pr_err("Failed to acquire DRC, rc: %d, drc index: %x\n", + rc, drc_index); + return -EINVAL; + } + + dn = dlpar_configure_connector(cpu_to_be32(drc_index), pmem_node); + if (!dn) { + pr_err("configure-connector failed for drc %x\n", drc_index); + dlpar_release_drc(drc_index); + return -EINVAL; + } + + /* NB: The of reconfig notifier creates platform device from the node */ + rc = dlpar_attach_node(dn, pmem_node); + if (rc) { + pr_err("Failed to attach node %pOF, rc: %d, drc index: %x\n", + dn, rc, drc_index); + + if (dlpar_release_drc(drc_index)) + dlpar_free_cc_nodes(dn); + + return rc; + } + + pr_info("Successfully added %pOF, drc index: %x\n", dn, drc_index); + + return 0; +} + +static ssize_t pmem_drc_remove_node(u32 drc_index) +{ + struct device_node *dn; + uint32_t index; + int rc; + + for_each_child_of_node(pmem_node, dn) { + if (of_property_read_u32(dn, "ibm,my-drc-index", &index)) + continue; + if (index == drc_index) + break; + } + + if (!dn) { + pr_err("Attempting to remove unused DRC index %x\n", drc_index); + return -ENODEV; + } + + pr_debug("Attempting to remove %pOF, drc index: %x\n", dn, drc_index); + + /* * NB: tears down the ibm,pmemory device as a side-effect */ + rc = dlpar_detach_node(dn); + if (rc) + return rc; + + rc = dlpar_release_drc(drc_index); + if (rc) { + pr_err("Failed to release drc (%x) for CPU %pOFn, rc: %d\n", + drc_index, dn, rc); + dlpar_attach_node(dn, pmem_node); + return rc; + } + + pr_info("Successfully removed PMEM with drc index: %x\n", drc_index); + + return 0; +} + +int dlpar_hp_pmem(struct pseries_hp_errorlog *hp_elog) +{ + u32 drc_index; + int rc; + + /* slim chance, but we might get a hotplug event while booting */ + if (!pmem_node) + pmem_node = of_find_node_by_type(NULL, "ibm,persistent-memory"); + if (!pmem_node) { + pr_err("Hotplug event for a pmem device, but none exists\n"); + return -ENODEV; + } + + if (hp_elog->id_type != PSERIES_HP_ELOG_ID_DRC_INDEX) { + pr_err("Unsupported hotplug event type %d\n", + hp_elog->id_type); + return -EINVAL; + } + + drc_index = hp_elog->_drc_u.drc_index; + + lock_device_hotplug(); + + if (hp_elog->action == PSERIES_HP_ELOG_ACTION_ADD) { + rc = pmem_drc_add_node(drc_index); + } else if (hp_elog->action == PSERIES_HP_ELOG_ACTION_REMOVE) { + rc = pmem_drc_remove_node(drc_index); + } else { + pr_err("Unsupported hotplug action (%d)\n", hp_elog->action); + rc = -EINVAL; + } + + unlock_device_hotplug(); + return rc; +} + +static const struct of_device_id drc_pmem_match[] = { + { .type = "ibm,persistent-memory", }, + {} +}; + +static int pseries_pmem_init(void) +{ + /* + * Only supported on POWER8 and above. + */ + if (!cpu_has_feature(CPU_FTR_ARCH_207S)) + return 0; + + pmem_node = of_find_node_by_type(NULL, "ibm,persistent-memory"); + if (!pmem_node) + return 0; + + /* + * The generic OF bus probe/populate handles creating platform devices + * from the child (ibm,pmemory) nodes. The generic code registers an of + * reconfig notifier to handle the hot-add/remove cases too. + */ + of_platform_bus_probe(pmem_node, drc_pmem_match, NULL); + + return 0; +} +machine_arch_initcall(pseries, pseries_pmem_init); diff --git a/arch/powerpc/platforms/pseries/power.c b/arch/powerpc/platforms/pseries/power.c new file mode 100644 index 000000000..3676cb297 --- /dev/null +++ b/arch/powerpc/platforms/pseries/power.c @@ -0,0 +1,72 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Interface for power-management for ppc64 compliant platform + * + * Manish Ahuja <mahuja@us.ibm.com> + * + * Feb 2007 + * + * Copyright (C) 2007 IBM Corporation. + */ + +#include <linux/kobject.h> +#include <linux/string.h> +#include <linux/errno.h> +#include <linux/init.h> +#include <asm/machdep.h> + +#include "pseries.h" + +unsigned long rtas_poweron_auto; /* default and normal state is 0 */ + +static ssize_t auto_poweron_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", rtas_poweron_auto); +} + +static ssize_t auto_poweron_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t n) +{ + int ret; + unsigned long ups_restart; + ret = sscanf(buf, "%lu", &ups_restart); + + if ((ret == 1) && ((ups_restart == 1) || (ups_restart == 0))){ + rtas_poweron_auto = ups_restart; + return n; + } + return -EINVAL; +} + +static struct kobj_attribute auto_poweron_attr = + __ATTR(auto_poweron, 0644, auto_poweron_show, auto_poweron_store); + +#ifndef CONFIG_PM +struct kobject *power_kobj; + +static struct attribute *g[] = { + &auto_poweron_attr.attr, + NULL, +}; + +static const struct attribute_group attr_group = { + .attrs = g, +}; + +static int __init pm_init(void) +{ + power_kobj = kobject_create_and_add("power", NULL); + if (!power_kobj) + return -ENOMEM; + return sysfs_create_group(power_kobj, &attr_group); +} +machine_core_initcall(pseries, pm_init); +#else +static int __init apo_pm_init(void) +{ + return (sysfs_create_file(power_kobj, &auto_poweron_attr.attr)); +} +machine_device_initcall(pseries, apo_pm_init); +#endif diff --git a/arch/powerpc/platforms/pseries/pseries.h b/arch/powerpc/platforms/pseries/pseries.h new file mode 100644 index 000000000..8376f03f9 --- /dev/null +++ b/arch/powerpc/platforms/pseries/pseries.h @@ -0,0 +1,131 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Copyright 2006 IBM Corporation. + */ + +#ifndef _PSERIES_PSERIES_H +#define _PSERIES_PSERIES_H + +#include <linux/interrupt.h> +#include <asm/rtas.h> + +struct device_node; + +void __init request_event_sources_irqs(struct device_node *np, + irq_handler_t handler, const char *name); + +#include <linux/of.h> + +struct pt_regs; + +extern int pSeries_system_reset_exception(struct pt_regs *regs); +extern int pSeries_machine_check_exception(struct pt_regs *regs); +extern long pseries_machine_check_realmode(struct pt_regs *regs); +void pSeries_machine_check_log_err(void); + +#ifdef CONFIG_SMP +extern void smp_init_pseries(void); + +/* Get state of physical CPU from query_cpu_stopped */ +int smp_query_cpu_stopped(unsigned int pcpu); +#define QCSS_STOPPED 0 +#define QCSS_STOPPING 1 +#define QCSS_NOT_STOPPED 2 +#define QCSS_HARDWARE_ERROR -1 +#define QCSS_HARDWARE_BUSY -2 +#else +static inline void smp_init_pseries(void) { } +#endif + +extern void pseries_kexec_cpu_down(int crash_shutdown, int secondary); +void pseries_machine_kexec(struct kimage *image); + +extern void pSeries_final_fixup(void); + +/* Poweron flag used for enabling auto ups restart */ +extern unsigned long rtas_poweron_auto; + +/* Dynamic logical Partitioning/Mobility */ +extern void dlpar_free_cc_nodes(struct device_node *); +extern void dlpar_free_cc_property(struct property *); +extern struct device_node *dlpar_configure_connector(__be32, + struct device_node *); +extern int dlpar_attach_node(struct device_node *, struct device_node *); +extern int dlpar_detach_node(struct device_node *); +extern int dlpar_acquire_drc(u32 drc_index); +extern int dlpar_release_drc(u32 drc_index); +extern int dlpar_unisolate_drc(u32 drc_index); + +void queue_hotplug_event(struct pseries_hp_errorlog *hp_errlog); +int handle_dlpar_errorlog(struct pseries_hp_errorlog *hp_errlog); + +#ifdef CONFIG_MEMORY_HOTPLUG +int dlpar_memory(struct pseries_hp_errorlog *hp_elog); +int dlpar_hp_pmem(struct pseries_hp_errorlog *hp_elog); +#else +static inline int dlpar_memory(struct pseries_hp_errorlog *hp_elog) +{ + return -EOPNOTSUPP; +} +static inline int dlpar_hp_pmem(struct pseries_hp_errorlog *hp_elog) +{ + return -EOPNOTSUPP; +} +#endif + +#ifdef CONFIG_HOTPLUG_CPU +int dlpar_cpu(struct pseries_hp_errorlog *hp_elog); +void pseries_cpu_hotplug_init(void); +#else +static inline int dlpar_cpu(struct pseries_hp_errorlog *hp_elog) +{ + return -EOPNOTSUPP; +} +static inline void pseries_cpu_hotplug_init(void) { } +#endif + +/* PCI root bridge prepare function override for pseries */ +struct pci_host_bridge; +int pseries_root_bridge_prepare(struct pci_host_bridge *bridge); + +extern struct pci_controller_ops pseries_pci_controller_ops; +int pseries_msi_allocate_domains(struct pci_controller *phb); +void pseries_msi_free_domains(struct pci_controller *phb); + +extern int CMO_PrPSP; +extern int CMO_SecPSP; +extern unsigned long CMO_PageSize; + +static inline int cmo_get_primary_psp(void) +{ + return CMO_PrPSP; +} + +static inline int cmo_get_secondary_psp(void) +{ + return CMO_SecPSP; +} + +static inline unsigned long cmo_get_page_size(void) +{ + return CMO_PageSize; +} + +int dlpar_workqueue_init(void); + +extern u32 pseries_security_flavor; +void pseries_setup_security_mitigations(void); + +#ifdef CONFIG_PPC_64S_HASH_MMU +void pseries_lpar_read_hblkrm_characteristics(void); +#else +static inline void pseries_lpar_read_hblkrm_characteristics(void) { } +#endif + +void pseries_rng_init(void); +#ifdef CONFIG_SPAPR_TCE_IOMMU +struct iommu_group *pSeries_pci_device_group(struct pci_controller *hose, + struct pci_dev *pdev); +#endif + +#endif /* _PSERIES_PSERIES_H */ diff --git a/arch/powerpc/platforms/pseries/pseries_energy.c b/arch/powerpc/platforms/pseries/pseries_energy.c new file mode 100644 index 000000000..2c661b798 --- /dev/null +++ b/arch/powerpc/platforms/pseries/pseries_energy.c @@ -0,0 +1,368 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * POWER platform energy management driver + * Copyright (C) 2010 IBM Corporation + * + * This pseries platform device driver provides access to + * platform energy management capabilities. + */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/errno.h> +#include <linux/init.h> +#include <linux/seq_file.h> +#include <linux/device.h> +#include <linux/cpu.h> +#include <linux/of.h> +#include <asm/cputhreads.h> +#include <asm/page.h> +#include <asm/hvcall.h> +#include <asm/firmware.h> +#include <asm/prom.h> + + +#define MODULE_VERS "1.0" +#define MODULE_NAME "pseries_energy" + +/* Driver flags */ + +static int sysfs_entries; + +/* Helper routines */ + +/* Helper Routines to convert between drc_index to cpu numbers */ + +static u32 cpu_to_drc_index(int cpu) +{ + struct device_node *dn = NULL; + struct property *info; + int thread_index; + int rc = 1; + u32 ret = 0; + + dn = of_find_node_by_path("/cpus"); + if (dn == NULL) + goto err; + + /* Convert logical cpu number to core number */ + thread_index = cpu_core_index_of_thread(cpu); + + info = of_find_property(dn, "ibm,drc-info", NULL); + if (info) { + struct of_drc_info drc; + int j; + u32 num_set_entries; + const __be32 *value; + + value = of_prop_next_u32(info, NULL, &num_set_entries); + if (!value) + goto err_of_node_put; + else + value++; + + for (j = 0; j < num_set_entries; j++) { + + of_read_drc_info_cell(&info, &value, &drc); + if (strncmp(drc.drc_type, "CPU", 3)) + goto err; + + if (thread_index < drc.last_drc_index) + break; + } + + ret = drc.drc_index_start + (thread_index * drc.sequential_inc); + } else { + u32 nr_drc_indexes, thread_drc_index; + + /* + * The first element of ibm,drc-indexes array is the + * number of drc_indexes returned in the list. Hence + * thread_index+1 will get the drc_index corresponding + * to core number thread_index. + */ + rc = of_property_read_u32_index(dn, "ibm,drc-indexes", + 0, &nr_drc_indexes); + if (rc) + goto err_of_node_put; + + WARN_ON_ONCE(thread_index > nr_drc_indexes); + rc = of_property_read_u32_index(dn, "ibm,drc-indexes", + thread_index + 1, + &thread_drc_index); + if (rc) + goto err_of_node_put; + + ret = thread_drc_index; + } + + rc = 0; + +err_of_node_put: + of_node_put(dn); +err: + if (rc) + printk(KERN_WARNING "cpu_to_drc_index(%d) failed", cpu); + return ret; +} + +static int drc_index_to_cpu(u32 drc_index) +{ + struct device_node *dn = NULL; + struct property *info; + const int *indexes; + int thread_index = 0, cpu = 0; + int rc = 1; + + dn = of_find_node_by_path("/cpus"); + if (dn == NULL) + goto err; + info = of_find_property(dn, "ibm,drc-info", NULL); + if (info) { + struct of_drc_info drc; + int j; + u32 num_set_entries; + const __be32 *value; + + value = of_prop_next_u32(info, NULL, &num_set_entries); + if (!value) + goto err_of_node_put; + else + value++; + + for (j = 0; j < num_set_entries; j++) { + + of_read_drc_info_cell(&info, &value, &drc); + if (strncmp(drc.drc_type, "CPU", 3)) + goto err; + + if (drc_index > drc.last_drc_index) { + cpu += drc.num_sequential_elems; + continue; + } + cpu += ((drc_index - drc.drc_index_start) / + drc.sequential_inc); + + thread_index = cpu_first_thread_of_core(cpu); + rc = 0; + break; + } + } else { + unsigned long int i; + + indexes = of_get_property(dn, "ibm,drc-indexes", NULL); + if (indexes == NULL) + goto err_of_node_put; + /* + * First element in the array is the number of drc_indexes + * returned. Search through the list to find the matching + * drc_index and get the core number + */ + for (i = 0; i < indexes[0]; i++) { + if (indexes[i + 1] == drc_index) + break; + } + /* Convert core number to logical cpu number */ + thread_index = cpu_first_thread_of_core(i); + rc = 0; + } + +err_of_node_put: + of_node_put(dn); +err: + if (rc) + printk(KERN_WARNING "drc_index_to_cpu(%d) failed", drc_index); + return thread_index; +} + +/* + * pseries hypervisor call H_BEST_ENERGY provides hints to OS on + * preferred logical cpus to activate or deactivate for optimized + * energy consumption. + */ + +#define FLAGS_MODE1 0x004E200000080E01UL +#define FLAGS_MODE2 0x004E200000080401UL +#define FLAGS_ACTIVATE 0x100 + +static ssize_t get_best_energy_list(char *page, int activate) +{ + int rc, cnt, i, cpu; + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE]; + unsigned long flags = 0; + u32 *buf_page; + char *s = page; + + buf_page = (u32 *) get_zeroed_page(GFP_KERNEL); + if (!buf_page) + return -ENOMEM; + + flags = FLAGS_MODE1; + if (activate) + flags |= FLAGS_ACTIVATE; + + rc = plpar_hcall9(H_BEST_ENERGY, retbuf, flags, 0, __pa(buf_page), + 0, 0, 0, 0, 0, 0); + if (rc != H_SUCCESS) { + free_page((unsigned long) buf_page); + return -EINVAL; + } + + cnt = retbuf[0]; + for (i = 0; i < cnt; i++) { + cpu = drc_index_to_cpu(buf_page[2*i+1]); + if ((cpu_online(cpu) && !activate) || + (!cpu_online(cpu) && activate)) + s += sprintf(s, "%d,", cpu); + } + if (s > page) { /* Something to show */ + s--; /* Suppress last comma */ + s += sprintf(s, "\n"); + } + + free_page((unsigned long) buf_page); + return s-page; +} + +static ssize_t get_best_energy_data(struct device *dev, + char *page, int activate) +{ + int rc; + unsigned long retbuf[PLPAR_HCALL9_BUFSIZE]; + unsigned long flags = 0; + + flags = FLAGS_MODE2; + if (activate) + flags |= FLAGS_ACTIVATE; + + rc = plpar_hcall9(H_BEST_ENERGY, retbuf, flags, + cpu_to_drc_index(dev->id), + 0, 0, 0, 0, 0, 0, 0); + + if (rc != H_SUCCESS) + return -EINVAL; + + return sprintf(page, "%lu\n", retbuf[1] >> 32); +} + +/* Wrapper functions */ + +static ssize_t cpu_activate_hint_list_show(struct device *dev, + struct device_attribute *attr, char *page) +{ + return get_best_energy_list(page, 1); +} + +static ssize_t cpu_deactivate_hint_list_show(struct device *dev, + struct device_attribute *attr, char *page) +{ + return get_best_energy_list(page, 0); +} + +static ssize_t percpu_activate_hint_show(struct device *dev, + struct device_attribute *attr, char *page) +{ + return get_best_energy_data(dev, page, 1); +} + +static ssize_t percpu_deactivate_hint_show(struct device *dev, + struct device_attribute *attr, char *page) +{ + return get_best_energy_data(dev, page, 0); +} + +/* + * Create sysfs interface: + * /sys/devices/system/cpu/pseries_activate_hint_list + * /sys/devices/system/cpu/pseries_deactivate_hint_list + * Comma separated list of cpus to activate or deactivate + * /sys/devices/system/cpu/cpuN/pseries_activate_hint + * /sys/devices/system/cpu/cpuN/pseries_deactivate_hint + * Per-cpu value of the hint + */ + +static struct device_attribute attr_cpu_activate_hint_list = + __ATTR(pseries_activate_hint_list, 0444, + cpu_activate_hint_list_show, NULL); + +static struct device_attribute attr_cpu_deactivate_hint_list = + __ATTR(pseries_deactivate_hint_list, 0444, + cpu_deactivate_hint_list_show, NULL); + +static struct device_attribute attr_percpu_activate_hint = + __ATTR(pseries_activate_hint, 0444, + percpu_activate_hint_show, NULL); + +static struct device_attribute attr_percpu_deactivate_hint = + __ATTR(pseries_deactivate_hint, 0444, + percpu_deactivate_hint_show, NULL); + +static int __init pseries_energy_init(void) +{ + int cpu, err; + struct device *cpu_dev, *dev_root; + + if (!firmware_has_feature(FW_FEATURE_BEST_ENERGY)) + return 0; /* H_BEST_ENERGY hcall not supported */ + + /* Create the sysfs files */ + dev_root = bus_get_dev_root(&cpu_subsys); + if (dev_root) { + err = device_create_file(dev_root, &attr_cpu_activate_hint_list); + if (!err) + err = device_create_file(dev_root, &attr_cpu_deactivate_hint_list); + put_device(dev_root); + if (err) + return err; + } + + for_each_possible_cpu(cpu) { + cpu_dev = get_cpu_device(cpu); + err = device_create_file(cpu_dev, + &attr_percpu_activate_hint); + if (err) + break; + err = device_create_file(cpu_dev, + &attr_percpu_deactivate_hint); + if (err) + break; + } + + if (err) + return err; + + sysfs_entries = 1; /* Removed entries on cleanup */ + return 0; + +} + +static void __exit pseries_energy_cleanup(void) +{ + int cpu; + struct device *cpu_dev, *dev_root; + + if (!sysfs_entries) + return; + + /* Remove the sysfs files */ + dev_root = bus_get_dev_root(&cpu_subsys); + if (dev_root) { + device_remove_file(dev_root, &attr_cpu_activate_hint_list); + device_remove_file(dev_root, &attr_cpu_deactivate_hint_list); + put_device(dev_root); + } + + for_each_possible_cpu(cpu) { + cpu_dev = get_cpu_device(cpu); + sysfs_remove_file(&cpu_dev->kobj, + &attr_percpu_activate_hint.attr); + sysfs_remove_file(&cpu_dev->kobj, + &attr_percpu_deactivate_hint.attr); + } +} + +module_init(pseries_energy_init); +module_exit(pseries_energy_cleanup); +MODULE_DESCRIPTION("Driver for pSeries platform energy management"); +MODULE_AUTHOR("Vaidyanathan Srinivasan"); +MODULE_LICENSE("GPL"); diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c new file mode 100644 index 000000000..adafd593d --- /dev/null +++ b/arch/powerpc/platforms/pseries/ras.c @@ -0,0 +1,882 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2001 Dave Engebretsen IBM Corporation + */ + +#include <linux/sched.h> +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/of.h> +#include <linux/fs.h> +#include <linux/reboot.h> +#include <linux/irq_work.h> + +#include <asm/machdep.h> +#include <asm/rtas.h> +#include <asm/firmware.h> +#include <asm/mce.h> + +#include "pseries.h" + +static unsigned char ras_log_buf[RTAS_ERROR_LOG_MAX]; +static DEFINE_SPINLOCK(ras_log_buf_lock); + +static int ras_check_exception_token; + +#define EPOW_SENSOR_TOKEN 9 +#define EPOW_SENSOR_INDEX 0 + +/* EPOW events counter variable */ +static int num_epow_events; + +static irqreturn_t ras_hotplug_interrupt(int irq, void *dev_id); +static irqreturn_t ras_epow_interrupt(int irq, void *dev_id); +static irqreturn_t ras_error_interrupt(int irq, void *dev_id); + +/* RTAS pseries MCE errorlog section. */ +struct pseries_mc_errorlog { + __be32 fru_id; + __be32 proc_id; + u8 error_type; + /* + * sub_err_type (1 byte). Bit fields depends on error_type + * + * MSB0 + * | + * V + * 01234567 + * XXXXXXXX + * + * For error_type == MC_ERROR_TYPE_UE + * XXXXXXXX + * X 1: Permanent or Transient UE. + * X 1: Effective address provided. + * X 1: Logical address provided. + * XX 2: Reserved. + * XXX 3: Type of UE error. + * + * For error_type == MC_ERROR_TYPE_SLB/ERAT/TLB + * XXXXXXXX + * X 1: Effective address provided. + * XXXXX 5: Reserved. + * XX 2: Type of SLB/ERAT/TLB error. + * + * For error_type == MC_ERROR_TYPE_CTRL_MEM_ACCESS + * XXXXXXXX + * X 1: Error causing address provided. + * XXX 3: Type of error. + * XXXX 4: Reserved. + */ + u8 sub_err_type; + u8 reserved_1[6]; + __be64 effective_address; + __be64 logical_address; +} __packed; + +/* RTAS pseries MCE error types */ +#define MC_ERROR_TYPE_UE 0x00 +#define MC_ERROR_TYPE_SLB 0x01 +#define MC_ERROR_TYPE_ERAT 0x02 +#define MC_ERROR_TYPE_UNKNOWN 0x03 +#define MC_ERROR_TYPE_TLB 0x04 +#define MC_ERROR_TYPE_D_CACHE 0x05 +#define MC_ERROR_TYPE_I_CACHE 0x07 +#define MC_ERROR_TYPE_CTRL_MEM_ACCESS 0x08 + +/* RTAS pseries MCE error sub types */ +#define MC_ERROR_UE_INDETERMINATE 0 +#define MC_ERROR_UE_IFETCH 1 +#define MC_ERROR_UE_PAGE_TABLE_WALK_IFETCH 2 +#define MC_ERROR_UE_LOAD_STORE 3 +#define MC_ERROR_UE_PAGE_TABLE_WALK_LOAD_STORE 4 + +#define UE_EFFECTIVE_ADDR_PROVIDED 0x40 +#define UE_LOGICAL_ADDR_PROVIDED 0x20 +#define MC_EFFECTIVE_ADDR_PROVIDED 0x80 + +#define MC_ERROR_SLB_PARITY 0 +#define MC_ERROR_SLB_MULTIHIT 1 +#define MC_ERROR_SLB_INDETERMINATE 2 + +#define MC_ERROR_ERAT_PARITY 1 +#define MC_ERROR_ERAT_MULTIHIT 2 +#define MC_ERROR_ERAT_INDETERMINATE 3 + +#define MC_ERROR_TLB_PARITY 1 +#define MC_ERROR_TLB_MULTIHIT 2 +#define MC_ERROR_TLB_INDETERMINATE 3 + +#define MC_ERROR_CTRL_MEM_ACCESS_PTABLE_WALK 0 +#define MC_ERROR_CTRL_MEM_ACCESS_OP_ACCESS 1 + +static inline u8 rtas_mc_error_sub_type(const struct pseries_mc_errorlog *mlog) +{ + switch (mlog->error_type) { + case MC_ERROR_TYPE_UE: + return (mlog->sub_err_type & 0x07); + case MC_ERROR_TYPE_SLB: + case MC_ERROR_TYPE_ERAT: + case MC_ERROR_TYPE_TLB: + return (mlog->sub_err_type & 0x03); + case MC_ERROR_TYPE_CTRL_MEM_ACCESS: + return (mlog->sub_err_type & 0x70) >> 4; + default: + return 0; + } +} + +/* + * Enable the hotplug interrupt late because processing them may touch other + * devices or systems (e.g. hugepages) that have not been initialized at the + * subsys stage. + */ +static int __init init_ras_hotplug_IRQ(void) +{ + struct device_node *np; + + /* Hotplug Events */ + np = of_find_node_by_path("/event-sources/hot-plug-events"); + if (np != NULL) { + if (dlpar_workqueue_init() == 0) + request_event_sources_irqs(np, ras_hotplug_interrupt, + "RAS_HOTPLUG"); + of_node_put(np); + } + + return 0; +} +machine_late_initcall(pseries, init_ras_hotplug_IRQ); + +/* + * Initialize handlers for the set of interrupts caused by hardware errors + * and power system events. + */ +static int __init init_ras_IRQ(void) +{ + struct device_node *np; + + ras_check_exception_token = rtas_function_token(RTAS_FN_CHECK_EXCEPTION); + + /* Internal Errors */ + np = of_find_node_by_path("/event-sources/internal-errors"); + if (np != NULL) { + request_event_sources_irqs(np, ras_error_interrupt, + "RAS_ERROR"); + of_node_put(np); + } + + /* EPOW Events */ + np = of_find_node_by_path("/event-sources/epow-events"); + if (np != NULL) { + request_event_sources_irqs(np, ras_epow_interrupt, "RAS_EPOW"); + of_node_put(np); + } + + return 0; +} +machine_subsys_initcall(pseries, init_ras_IRQ); + +#define EPOW_SHUTDOWN_NORMAL 1 +#define EPOW_SHUTDOWN_ON_UPS 2 +#define EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS 3 +#define EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH 4 + +static void handle_system_shutdown(char event_modifier) +{ + switch (event_modifier) { + case EPOW_SHUTDOWN_NORMAL: + pr_emerg("Power off requested\n"); + orderly_poweroff(true); + break; + + case EPOW_SHUTDOWN_ON_UPS: + pr_emerg("Loss of system power detected. System is running on" + " UPS/battery. Check RTAS error log for details\n"); + break; + + case EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS: + pr_emerg("Loss of system critical functions detected. Check" + " RTAS error log for details\n"); + orderly_poweroff(true); + break; + + case EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH: + pr_emerg("High ambient temperature detected. Check RTAS" + " error log for details\n"); + orderly_poweroff(true); + break; + + default: + pr_err("Unknown power/cooling shutdown event (modifier = %d)\n", + event_modifier); + } +} + +struct epow_errorlog { + unsigned char sensor_value; + unsigned char event_modifier; + unsigned char extended_modifier; + unsigned char reserved; + unsigned char platform_reason; +}; + +#define EPOW_RESET 0 +#define EPOW_WARN_COOLING 1 +#define EPOW_WARN_POWER 2 +#define EPOW_SYSTEM_SHUTDOWN 3 +#define EPOW_SYSTEM_HALT 4 +#define EPOW_MAIN_ENCLOSURE 5 +#define EPOW_POWER_OFF 7 + +static void rtas_parse_epow_errlog(struct rtas_error_log *log) +{ + struct pseries_errorlog *pseries_log; + struct epow_errorlog *epow_log; + char action_code; + char modifier; + + pseries_log = get_pseries_errorlog(log, PSERIES_ELOG_SECT_ID_EPOW); + if (pseries_log == NULL) + return; + + epow_log = (struct epow_errorlog *)pseries_log->data; + action_code = epow_log->sensor_value & 0xF; /* bottom 4 bits */ + modifier = epow_log->event_modifier & 0xF; /* bottom 4 bits */ + + switch (action_code) { + case EPOW_RESET: + if (num_epow_events) { + pr_info("Non critical power/cooling issue cleared\n"); + num_epow_events--; + } + break; + + case EPOW_WARN_COOLING: + pr_info("Non-critical cooling issue detected. Check RTAS error" + " log for details\n"); + break; + + case EPOW_WARN_POWER: + pr_info("Non-critical power issue detected. Check RTAS error" + " log for details\n"); + break; + + case EPOW_SYSTEM_SHUTDOWN: + handle_system_shutdown(modifier); + break; + + case EPOW_SYSTEM_HALT: + pr_emerg("Critical power/cooling issue detected. Check RTAS" + " error log for details. Powering off.\n"); + orderly_poweroff(true); + break; + + case EPOW_MAIN_ENCLOSURE: + case EPOW_POWER_OFF: + pr_emerg("System about to lose power. Check RTAS error log " + " for details. Powering off immediately.\n"); + emergency_sync(); + kernel_power_off(); + break; + + default: + pr_err("Unknown power/cooling event (action code = %d)\n", + action_code); + } + + /* Increment epow events counter variable */ + if (action_code != EPOW_RESET) + num_epow_events++; +} + +static irqreturn_t ras_hotplug_interrupt(int irq, void *dev_id) +{ + struct pseries_errorlog *pseries_log; + struct pseries_hp_errorlog *hp_elog; + + spin_lock(&ras_log_buf_lock); + + rtas_call(ras_check_exception_token, 6, 1, NULL, + RTAS_VECTOR_EXTERNAL_INTERRUPT, virq_to_hw(irq), + RTAS_HOTPLUG_EVENTS, 0, __pa(&ras_log_buf), + rtas_get_error_log_max()); + + pseries_log = get_pseries_errorlog((struct rtas_error_log *)ras_log_buf, + PSERIES_ELOG_SECT_ID_HOTPLUG); + hp_elog = (struct pseries_hp_errorlog *)pseries_log->data; + + /* + * Since PCI hotplug is not currently supported on pseries, put PCI + * hotplug events on the ras_log_buf to be handled by rtas_errd. + */ + if (hp_elog->resource == PSERIES_HP_ELOG_RESOURCE_MEM || + hp_elog->resource == PSERIES_HP_ELOG_RESOURCE_CPU || + hp_elog->resource == PSERIES_HP_ELOG_RESOURCE_PMEM) + queue_hotplug_event(hp_elog); + else + log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, 0); + + spin_unlock(&ras_log_buf_lock); + return IRQ_HANDLED; +} + +/* Handle environmental and power warning (EPOW) interrupts. */ +static irqreturn_t ras_epow_interrupt(int irq, void *dev_id) +{ + int state; + int critical; + + rtas_get_sensor_fast(EPOW_SENSOR_TOKEN, EPOW_SENSOR_INDEX, &state); + + if (state > 3) + critical = 1; /* Time Critical */ + else + critical = 0; + + spin_lock(&ras_log_buf_lock); + + rtas_call(ras_check_exception_token, 6, 1, NULL, RTAS_VECTOR_EXTERNAL_INTERRUPT, + virq_to_hw(irq), RTAS_EPOW_WARNING, critical, __pa(&ras_log_buf), + rtas_get_error_log_max()); + + log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, 0); + + rtas_parse_epow_errlog((struct rtas_error_log *)ras_log_buf); + + spin_unlock(&ras_log_buf_lock); + return IRQ_HANDLED; +} + +/* + * Handle hardware error interrupts. + * + * RTAS check-exception is called to collect data on the exception. If + * the error is deemed recoverable, we log a warning and return. + * For nonrecoverable errors, an error is logged and we stop all processing + * as quickly as possible in order to prevent propagation of the failure. + */ +static irqreturn_t ras_error_interrupt(int irq, void *dev_id) +{ + struct rtas_error_log *rtas_elog; + int status; + int fatal; + + spin_lock(&ras_log_buf_lock); + + status = rtas_call(ras_check_exception_token, 6, 1, NULL, + RTAS_VECTOR_EXTERNAL_INTERRUPT, + virq_to_hw(irq), + RTAS_INTERNAL_ERROR, 1 /* Time Critical */, + __pa(&ras_log_buf), + rtas_get_error_log_max()); + + rtas_elog = (struct rtas_error_log *)ras_log_buf; + + if (status == 0 && + rtas_error_severity(rtas_elog) >= RTAS_SEVERITY_ERROR_SYNC) + fatal = 1; + else + fatal = 0; + + /* format and print the extended information */ + log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, fatal); + + if (fatal) { + pr_emerg("Fatal hardware error detected. Check RTAS error" + " log for details. Powering off immediately\n"); + emergency_sync(); + kernel_power_off(); + } else { + pr_err("Recoverable hardware error detected\n"); + } + + spin_unlock(&ras_log_buf_lock); + return IRQ_HANDLED; +} + +/* + * Some versions of FWNMI place the buffer inside the 4kB page starting at + * 0x7000. Other versions place it inside the rtas buffer. We check both. + * Minimum size of the buffer is 16 bytes. + */ +#define VALID_FWNMI_BUFFER(A) \ + ((((A) >= 0x7000) && ((A) <= 0x8000 - 16)) || \ + (((A) >= rtas.base) && ((A) <= (rtas.base + rtas.size - 16)))) + +static inline struct rtas_error_log *fwnmi_get_errlog(void) +{ + return (struct rtas_error_log *)local_paca->mce_data_buf; +} + +static __be64 *fwnmi_get_savep(struct pt_regs *regs) +{ + unsigned long savep_ra; + + /* Mask top two bits */ + savep_ra = regs->gpr[3] & ~(0x3UL << 62); + if (!VALID_FWNMI_BUFFER(savep_ra)) { + printk(KERN_ERR "FWNMI: corrupt r3 0x%016lx\n", regs->gpr[3]); + return NULL; + } + + return __va(savep_ra); +} + +/* + * Get the error information for errors coming through the + * FWNMI vectors. The pt_regs' r3 will be updated to reflect + * the actual r3 if possible, and a ptr to the error log entry + * will be returned if found. + * + * Use one buffer mce_data_buf per cpu to store RTAS error. + * + * The mce_data_buf does not have any locks or protection around it, + * if a second machine check comes in, or a system reset is done + * before we have logged the error, then we will get corruption in the + * error log. This is preferable over holding off on calling + * ibm,nmi-interlock which would result in us checkstopping if a + * second machine check did come in. + */ +static struct rtas_error_log *fwnmi_get_errinfo(struct pt_regs *regs) +{ + struct rtas_error_log *h; + __be64 *savep; + + savep = fwnmi_get_savep(regs); + if (!savep) + return NULL; + + regs->gpr[3] = be64_to_cpu(savep[0]); /* restore original r3 */ + + h = (struct rtas_error_log *)&savep[1]; + /* Use the per cpu buffer from paca to store rtas error log */ + memset(local_paca->mce_data_buf, 0, RTAS_ERROR_LOG_MAX); + if (!rtas_error_extended(h)) { + memcpy(local_paca->mce_data_buf, h, sizeof(__u64)); + } else { + int len, error_log_length; + + error_log_length = 8 + rtas_error_extended_log_length(h); + len = min_t(int, error_log_length, RTAS_ERROR_LOG_MAX); + memcpy(local_paca->mce_data_buf, h, len); + } + + return (struct rtas_error_log *)local_paca->mce_data_buf; +} + +/* Call this when done with the data returned by FWNMI_get_errinfo. + * It will release the saved data area for other CPUs in the + * partition to receive FWNMI errors. + */ +static void fwnmi_release_errinfo(void) +{ + struct rtas_args rtas_args; + int ret; + + /* + * On pseries, the machine check stack is limited to under 4GB, so + * args can be on-stack. + */ + rtas_call_unlocked(&rtas_args, ibm_nmi_interlock_token, 0, 1, NULL); + ret = be32_to_cpu(rtas_args.rets[0]); + if (ret != 0) + printk(KERN_ERR "FWNMI: nmi-interlock failed: %d\n", ret); +} + +int pSeries_system_reset_exception(struct pt_regs *regs) +{ +#ifdef __LITTLE_ENDIAN__ + /* + * Some firmware byteswaps SRR registers and gives incorrect SRR1. Try + * to detect the bad SRR1 pattern here. Flip the NIP back to correct + * endian for reporting purposes. Unfortunately the MSR can't be fixed, + * so clear it. It will be missing MSR_RI so we won't try to recover. + */ + if ((be64_to_cpu(regs->msr) & + (MSR_LE|MSR_RI|MSR_DR|MSR_IR|MSR_ME|MSR_PR| + MSR_ILE|MSR_HV|MSR_SF)) == (MSR_DR|MSR_SF)) { + regs_set_return_ip(regs, be64_to_cpu((__be64)regs->nip)); + regs_set_return_msr(regs, 0); + } +#endif + + if (fwnmi_active) { + __be64 *savep; + + /* + * Firmware (PowerVM and KVM) saves r3 to a save area like + * machine check, which is not exactly what PAPR (2.9) + * suggests but there is no way to detect otherwise, so this + * is the interface now. + * + * System resets do not save any error log or require an + * "ibm,nmi-interlock" rtas call to release. + */ + + savep = fwnmi_get_savep(regs); + if (savep) + regs->gpr[3] = be64_to_cpu(savep[0]); /* restore original r3 */ + } + + if (smp_handle_nmi_ipi(regs)) + return 1; + + return 0; /* need to perform reset */ +} + +static int mce_handle_err_realmode(int disposition, u8 error_type) +{ +#ifdef CONFIG_PPC_BOOK3S_64 + if (disposition == RTAS_DISP_NOT_RECOVERED) { + switch (error_type) { + case MC_ERROR_TYPE_ERAT: + flush_erat(); + disposition = RTAS_DISP_FULLY_RECOVERED; + break; + case MC_ERROR_TYPE_SLB: +#ifdef CONFIG_PPC_64S_HASH_MMU + /* + * Store the old slb content in paca before flushing. + * Print this when we go to virtual mode. + * There are chances that we may hit MCE again if there + * is a parity error on the SLB entry we trying to read + * for saving. Hence limit the slb saving to single + * level of recursion. + */ + if (local_paca->in_mce == 1) + slb_save_contents(local_paca->mce_faulty_slbs); + flush_and_reload_slb(); + disposition = RTAS_DISP_FULLY_RECOVERED; +#endif + break; + default: + break; + } + } else if (disposition == RTAS_DISP_LIMITED_RECOVERY) { + /* Platform corrected itself but could be degraded */ + pr_err("MCE: limited recovery, system may be degraded\n"); + disposition = RTAS_DISP_FULLY_RECOVERED; + } +#endif + return disposition; +} + +static int mce_handle_err_virtmode(struct pt_regs *regs, + struct rtas_error_log *errp, + struct pseries_mc_errorlog *mce_log, + int disposition) +{ + struct mce_error_info mce_err = { 0 }; + int initiator = rtas_error_initiator(errp); + int severity = rtas_error_severity(errp); + unsigned long eaddr = 0, paddr = 0; + u8 error_type, err_sub_type; + + if (!mce_log) + goto out; + + error_type = mce_log->error_type; + err_sub_type = rtas_mc_error_sub_type(mce_log); + + if (initiator == RTAS_INITIATOR_UNKNOWN) + mce_err.initiator = MCE_INITIATOR_UNKNOWN; + else if (initiator == RTAS_INITIATOR_CPU) + mce_err.initiator = MCE_INITIATOR_CPU; + else if (initiator == RTAS_INITIATOR_PCI) + mce_err.initiator = MCE_INITIATOR_PCI; + else if (initiator == RTAS_INITIATOR_ISA) + mce_err.initiator = MCE_INITIATOR_ISA; + else if (initiator == RTAS_INITIATOR_MEMORY) + mce_err.initiator = MCE_INITIATOR_MEMORY; + else if (initiator == RTAS_INITIATOR_POWERMGM) + mce_err.initiator = MCE_INITIATOR_POWERMGM; + else + mce_err.initiator = MCE_INITIATOR_UNKNOWN; + + if (severity == RTAS_SEVERITY_NO_ERROR) + mce_err.severity = MCE_SEV_NO_ERROR; + else if (severity == RTAS_SEVERITY_EVENT) + mce_err.severity = MCE_SEV_WARNING; + else if (severity == RTAS_SEVERITY_WARNING) + mce_err.severity = MCE_SEV_WARNING; + else if (severity == RTAS_SEVERITY_ERROR_SYNC) + mce_err.severity = MCE_SEV_SEVERE; + else if (severity == RTAS_SEVERITY_ERROR) + mce_err.severity = MCE_SEV_SEVERE; + else + mce_err.severity = MCE_SEV_FATAL; + + if (severity <= RTAS_SEVERITY_ERROR_SYNC) + mce_err.sync_error = true; + else + mce_err.sync_error = false; + + mce_err.error_type = MCE_ERROR_TYPE_UNKNOWN; + mce_err.error_class = MCE_ECLASS_UNKNOWN; + + switch (error_type) { + case MC_ERROR_TYPE_UE: + mce_err.error_type = MCE_ERROR_TYPE_UE; + mce_common_process_ue(regs, &mce_err); + if (mce_err.ignore_event) + disposition = RTAS_DISP_FULLY_RECOVERED; + switch (err_sub_type) { + case MC_ERROR_UE_IFETCH: + mce_err.u.ue_error_type = MCE_UE_ERROR_IFETCH; + break; + case MC_ERROR_UE_PAGE_TABLE_WALK_IFETCH: + mce_err.u.ue_error_type = MCE_UE_ERROR_PAGE_TABLE_WALK_IFETCH; + break; + case MC_ERROR_UE_LOAD_STORE: + mce_err.u.ue_error_type = MCE_UE_ERROR_LOAD_STORE; + break; + case MC_ERROR_UE_PAGE_TABLE_WALK_LOAD_STORE: + mce_err.u.ue_error_type = MCE_UE_ERROR_PAGE_TABLE_WALK_LOAD_STORE; + break; + case MC_ERROR_UE_INDETERMINATE: + default: + mce_err.u.ue_error_type = MCE_UE_ERROR_INDETERMINATE; + break; + } + if (mce_log->sub_err_type & UE_EFFECTIVE_ADDR_PROVIDED) + eaddr = be64_to_cpu(mce_log->effective_address); + + if (mce_log->sub_err_type & UE_LOGICAL_ADDR_PROVIDED) { + paddr = be64_to_cpu(mce_log->logical_address); + } else if (mce_log->sub_err_type & UE_EFFECTIVE_ADDR_PROVIDED) { + unsigned long pfn; + + pfn = addr_to_pfn(regs, eaddr); + if (pfn != ULONG_MAX) + paddr = pfn << PAGE_SHIFT; + } + + break; + case MC_ERROR_TYPE_SLB: + mce_err.error_type = MCE_ERROR_TYPE_SLB; + switch (err_sub_type) { + case MC_ERROR_SLB_PARITY: + mce_err.u.slb_error_type = MCE_SLB_ERROR_PARITY; + break; + case MC_ERROR_SLB_MULTIHIT: + mce_err.u.slb_error_type = MCE_SLB_ERROR_MULTIHIT; + break; + case MC_ERROR_SLB_INDETERMINATE: + default: + mce_err.u.slb_error_type = MCE_SLB_ERROR_INDETERMINATE; + break; + } + if (mce_log->sub_err_type & MC_EFFECTIVE_ADDR_PROVIDED) + eaddr = be64_to_cpu(mce_log->effective_address); + break; + case MC_ERROR_TYPE_ERAT: + mce_err.error_type = MCE_ERROR_TYPE_ERAT; + switch (err_sub_type) { + case MC_ERROR_ERAT_PARITY: + mce_err.u.erat_error_type = MCE_ERAT_ERROR_PARITY; + break; + case MC_ERROR_ERAT_MULTIHIT: + mce_err.u.erat_error_type = MCE_ERAT_ERROR_MULTIHIT; + break; + case MC_ERROR_ERAT_INDETERMINATE: + default: + mce_err.u.erat_error_type = MCE_ERAT_ERROR_INDETERMINATE; + break; + } + if (mce_log->sub_err_type & MC_EFFECTIVE_ADDR_PROVIDED) + eaddr = be64_to_cpu(mce_log->effective_address); + break; + case MC_ERROR_TYPE_TLB: + mce_err.error_type = MCE_ERROR_TYPE_TLB; + switch (err_sub_type) { + case MC_ERROR_TLB_PARITY: + mce_err.u.tlb_error_type = MCE_TLB_ERROR_PARITY; + break; + case MC_ERROR_TLB_MULTIHIT: + mce_err.u.tlb_error_type = MCE_TLB_ERROR_MULTIHIT; + break; + case MC_ERROR_TLB_INDETERMINATE: + default: + mce_err.u.tlb_error_type = MCE_TLB_ERROR_INDETERMINATE; + break; + } + if (mce_log->sub_err_type & MC_EFFECTIVE_ADDR_PROVIDED) + eaddr = be64_to_cpu(mce_log->effective_address); + break; + case MC_ERROR_TYPE_D_CACHE: + mce_err.error_type = MCE_ERROR_TYPE_DCACHE; + break; + case MC_ERROR_TYPE_I_CACHE: + mce_err.error_type = MCE_ERROR_TYPE_ICACHE; + break; + case MC_ERROR_TYPE_CTRL_MEM_ACCESS: + mce_err.error_type = MCE_ERROR_TYPE_RA; + switch (err_sub_type) { + case MC_ERROR_CTRL_MEM_ACCESS_PTABLE_WALK: + mce_err.u.ra_error_type = + MCE_RA_ERROR_PAGE_TABLE_WALK_LOAD_STORE_FOREIGN; + break; + case MC_ERROR_CTRL_MEM_ACCESS_OP_ACCESS: + mce_err.u.ra_error_type = + MCE_RA_ERROR_LOAD_STORE_FOREIGN; + break; + } + if (mce_log->sub_err_type & MC_EFFECTIVE_ADDR_PROVIDED) + eaddr = be64_to_cpu(mce_log->effective_address); + break; + case MC_ERROR_TYPE_UNKNOWN: + default: + mce_err.error_type = MCE_ERROR_TYPE_UNKNOWN; + break; + } +out: + save_mce_event(regs, disposition == RTAS_DISP_FULLY_RECOVERED, + &mce_err, regs->nip, eaddr, paddr); + return disposition; +} + +static int mce_handle_error(struct pt_regs *regs, struct rtas_error_log *errp) +{ + struct pseries_errorlog *pseries_log; + struct pseries_mc_errorlog *mce_log = NULL; + int disposition = rtas_error_disposition(errp); + u8 error_type; + + if (!rtas_error_extended(errp)) + goto out; + + pseries_log = get_pseries_errorlog(errp, PSERIES_ELOG_SECT_ID_MCE); + if (!pseries_log) + goto out; + + mce_log = (struct pseries_mc_errorlog *)pseries_log->data; + error_type = mce_log->error_type; + + disposition = mce_handle_err_realmode(disposition, error_type); +out: + disposition = mce_handle_err_virtmode(regs, errp, mce_log, + disposition); + return disposition; +} + +/* + * Process MCE rtas errlog event. + */ +void pSeries_machine_check_log_err(void) +{ + struct rtas_error_log *err; + + err = fwnmi_get_errlog(); + log_error((char *)err, ERR_TYPE_RTAS_LOG, 0); +} + +/* + * See if we can recover from a machine check exception. + * This is only called on power4 (or above) and only via + * the Firmware Non-Maskable Interrupts (fwnmi) handler + * which provides the error analysis for us. + * + * Return 1 if corrected (or delivered a signal). + * Return 0 if there is nothing we can do. + */ +static int recover_mce(struct pt_regs *regs, struct machine_check_event *evt) +{ + int recovered = 0; + + if (regs_is_unrecoverable(regs)) { + /* If MSR_RI isn't set, we cannot recover */ + pr_err("Machine check interrupt unrecoverable: MSR(RI=0)\n"); + recovered = 0; + } else if (evt->disposition == MCE_DISPOSITION_RECOVERED) { + /* Platform corrected itself */ + recovered = 1; + } else if (evt->severity == MCE_SEV_FATAL) { + /* Fatal machine check */ + pr_err("Machine check interrupt is fatal\n"); + recovered = 0; + } + + if (!recovered && evt->sync_error) { + /* + * Try to kill processes if we get a synchronous machine check + * (e.g., one caused by execution of this instruction). This + * will devolve into a panic if we try to kill init or are in + * an interrupt etc. + * + * TODO: Queue up this address for hwpoisioning later. + * TODO: This is not quite right for d-side machine + * checks ->nip is not necessarily the important + * address. + */ + if ((user_mode(regs))) { + _exception(SIGBUS, regs, BUS_MCEERR_AR, regs->nip); + recovered = 1; + } else if (die_will_crash()) { + /* + * die() would kill the kernel, so better to go via + * the platform reboot code that will log the + * machine check. + */ + recovered = 0; + } else { + die_mce("Machine check", regs, SIGBUS); + recovered = 1; + } + } + + return recovered; +} + +/* + * Handle a machine check. + * + * Note that on Power 4 and beyond Firmware Non-Maskable Interrupts (fwnmi) + * should be present. If so the handler which called us tells us if the + * error was recovered (never true if RI=0). + * + * On hardware prior to Power 4 these exceptions were asynchronous which + * means we can't tell exactly where it occurred and so we can't recover. + */ +int pSeries_machine_check_exception(struct pt_regs *regs) +{ + struct machine_check_event evt; + + if (!get_mce_event(&evt, MCE_EVENT_RELEASE)) + return 0; + + /* Print things out */ + if (evt.version != MCE_V1) { + pr_err("Machine Check Exception, Unknown event version %d !\n", + evt.version); + return 0; + } + machine_check_print_event_info(&evt, user_mode(regs), false); + + if (recover_mce(regs, &evt)) + return 1; + + return 0; +} + +long pseries_machine_check_realmode(struct pt_regs *regs) +{ + struct rtas_error_log *errp; + int disposition; + + if (fwnmi_active) { + errp = fwnmi_get_errinfo(regs); + /* + * Call to fwnmi_release_errinfo() in real mode causes kernel + * to panic. Hence we will call it as soon as we go into + * virtual mode. + */ + disposition = mce_handle_error(regs, errp); + + fwnmi_release_errinfo(); + + if (disposition == RTAS_DISP_FULLY_RECOVERED) + return 1; + } + + return 0; +} diff --git a/arch/powerpc/platforms/pseries/reconfig.c b/arch/powerpc/platforms/pseries/reconfig.c new file mode 100644 index 000000000..599bd2c78 --- /dev/null +++ b/arch/powerpc/platforms/pseries/reconfig.c @@ -0,0 +1,414 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * pSeries_reconfig.c - support for dynamic reconfiguration (including PCI + * Hotplug and Dynamic Logical Partitioning on RPA platforms). + * + * Copyright (C) 2005 Nathan Lynch + * Copyright (C) 2005 IBM Corporation + */ + +#include <linux/kernel.h> +#include <linux/notifier.h> +#include <linux/proc_fs.h> +#include <linux/security.h> +#include <linux/slab.h> +#include <linux/of.h> + +#include <asm/machdep.h> +#include <linux/uaccess.h> +#include <asm/mmu.h> + +#include "of_helpers.h" + +static int pSeries_reconfig_add_node(const char *path, struct property *proplist) +{ + struct device_node *np; + int err = -ENOMEM; + + np = kzalloc(sizeof(*np), GFP_KERNEL); + if (!np) + goto out_err; + + np->full_name = kstrdup(kbasename(path), GFP_KERNEL); + if (!np->full_name) + goto out_err; + + np->properties = proplist; + of_node_set_flag(np, OF_DYNAMIC); + of_node_init(np); + + np->parent = pseries_of_derive_parent(path); + if (IS_ERR(np->parent)) { + err = PTR_ERR(np->parent); + goto out_err; + } + + err = of_attach_node(np); + if (err) { + printk(KERN_ERR "Failed to add device node %s\n", path); + goto out_err; + } + + of_node_put(np->parent); + + return 0; + +out_err: + if (np) { + of_node_put(np->parent); + kfree(np->full_name); + kfree(np); + } + return err; +} + +static int pSeries_reconfig_remove_node(struct device_node *np) +{ + struct device_node *parent, *child; + + parent = of_get_parent(np); + if (!parent) + return -EINVAL; + + if ((child = of_get_next_child(np, NULL))) { + of_node_put(child); + of_node_put(parent); + return -EBUSY; + } + + of_detach_node(np); + of_node_put(parent); + return 0; +} + +/* + * /proc/powerpc/ofdt - yucky binary interface for adding and removing + * OF device nodes. Should be deprecated as soon as we get an + * in-kernel wrapper for the RTAS ibm,configure-connector call. + */ + +static void release_prop_list(const struct property *prop) +{ + struct property *next; + for (; prop; prop = next) { + next = prop->next; + kfree(prop->name); + kfree(prop->value); + kfree(prop); + } + +} + +/** + * parse_next_property - process the next property from raw input buffer + * @buf: input buffer, must be nul-terminated + * @end: end of the input buffer + 1, for validation + * @name: return value; set to property name in buf + * @length: return value; set to length of value + * @value: return value; set to the property value in buf + * + * Note that the caller must make copies of the name and value returned, + * this function does no allocation or copying of the data. Return value + * is set to the next name in buf, or NULL on error. + */ +static char * parse_next_property(char *buf, char *end, char **name, int *length, + unsigned char **value) +{ + char *tmp; + + *name = buf; + + tmp = strchr(buf, ' '); + if (!tmp) { + printk(KERN_ERR "property parse failed in %s at line %d\n", + __func__, __LINE__); + return NULL; + } + *tmp = '\0'; + + if (++tmp >= end) { + printk(KERN_ERR "property parse failed in %s at line %d\n", + __func__, __LINE__); + return NULL; + } + + /* now we're on the length */ + *length = -1; + *length = simple_strtoul(tmp, &tmp, 10); + if (*length == -1) { + printk(KERN_ERR "property parse failed in %s at line %d\n", + __func__, __LINE__); + return NULL; + } + if (*tmp != ' ' || ++tmp >= end) { + printk(KERN_ERR "property parse failed in %s at line %d\n", + __func__, __LINE__); + return NULL; + } + + /* now we're on the value */ + *value = tmp; + tmp += *length; + if (tmp > end) { + printk(KERN_ERR "property parse failed in %s at line %d\n", + __func__, __LINE__); + return NULL; + } + else if (tmp < end && *tmp != ' ' && *tmp != '\0') { + printk(KERN_ERR "property parse failed in %s at line %d\n", + __func__, __LINE__); + return NULL; + } + tmp++; + + /* and now we should be on the next name, or the end */ + return tmp; +} + +static struct property *new_property(const char *name, const int length, + const unsigned char *value, struct property *last) +{ + struct property *new = kzalloc(sizeof(*new), GFP_KERNEL); + + if (!new) + return NULL; + + if (!(new->name = kstrdup(name, GFP_KERNEL))) + goto cleanup; + if (!(new->value = kmalloc(length + 1, GFP_KERNEL))) + goto cleanup; + + memcpy(new->value, value, length); + *(((char *)new->value) + length) = 0; + new->length = length; + new->next = last; + return new; + +cleanup: + kfree(new->name); + kfree(new->value); + kfree(new); + return NULL; +} + +static int do_add_node(char *buf, size_t bufsize) +{ + char *path, *end, *name; + struct device_node *np; + struct property *prop = NULL; + unsigned char* value; + int length, rv = 0; + + end = buf + bufsize; + path = buf; + buf = strchr(buf, ' '); + if (!buf) + return -EINVAL; + *buf = '\0'; + buf++; + + if ((np = of_find_node_by_path(path))) { + of_node_put(np); + return -EINVAL; + } + + /* rv = build_prop_list(tmp, bufsize - (tmp - buf), &proplist); */ + while (buf < end && + (buf = parse_next_property(buf, end, &name, &length, &value))) { + struct property *last = prop; + + prop = new_property(name, length, value, last); + if (!prop) { + rv = -ENOMEM; + prop = last; + goto out; + } + } + if (!buf) { + rv = -EINVAL; + goto out; + } + + rv = pSeries_reconfig_add_node(path, prop); + +out: + if (rv) + release_prop_list(prop); + return rv; +} + +static int do_remove_node(char *buf) +{ + struct device_node *node; + int rv = -ENODEV; + + if ((node = of_find_node_by_path(buf))) + rv = pSeries_reconfig_remove_node(node); + + of_node_put(node); + return rv; +} + +static char *parse_node(char *buf, size_t bufsize, struct device_node **npp) +{ + char *handle_str; + phandle handle; + *npp = NULL; + + handle_str = buf; + + buf = strchr(buf, ' '); + if (!buf) + return NULL; + *buf = '\0'; + buf++; + + handle = simple_strtoul(handle_str, NULL, 0); + + *npp = of_find_node_by_phandle(handle); + return buf; +} + +static int do_add_property(char *buf, size_t bufsize) +{ + struct property *prop = NULL; + struct device_node *np; + unsigned char *value; + char *name, *end; + int length; + end = buf + bufsize; + buf = parse_node(buf, bufsize, &np); + + if (!np) + return -ENODEV; + + if (parse_next_property(buf, end, &name, &length, &value) == NULL) + return -EINVAL; + + prop = new_property(name, length, value, NULL); + if (!prop) + return -ENOMEM; + + of_add_property(np, prop); + + return 0; +} + +static int do_remove_property(char *buf, size_t bufsize) +{ + struct device_node *np; + char *tmp; + buf = parse_node(buf, bufsize, &np); + + if (!np) + return -ENODEV; + + tmp = strchr(buf,' '); + if (tmp) + *tmp = '\0'; + + if (strlen(buf) == 0) + return -EINVAL; + + return of_remove_property(np, of_find_property(np, buf, NULL)); +} + +static int do_update_property(char *buf, size_t bufsize) +{ + struct device_node *np; + unsigned char *value; + char *name, *end, *next_prop; + int length; + struct property *newprop; + buf = parse_node(buf, bufsize, &np); + end = buf + bufsize; + + if (!np) + return -ENODEV; + + next_prop = parse_next_property(buf, end, &name, &length, &value); + if (!next_prop) + return -EINVAL; + + if (!strlen(name)) + return -ENODEV; + + newprop = new_property(name, length, value, NULL); + if (!newprop) + return -ENOMEM; + + if (!strcmp(name, "slb-size") || !strcmp(name, "ibm,slb-size")) + slb_set_size(*(int *)value); + + return of_update_property(np, newprop); +} + +/** + * ofdt_write - perform operations on the Open Firmware device tree + * + * @file: not used + * @buf: command and arguments + * @count: size of the command buffer + * @off: not used + * + * Operations supported at this time are addition and removal of + * whole nodes along with their properties. Operations on individual + * properties are not implemented (yet). + */ +static ssize_t ofdt_write(struct file *file, const char __user *buf, size_t count, + loff_t *off) +{ + int rv; + char *kbuf; + char *tmp; + + rv = security_locked_down(LOCKDOWN_DEVICE_TREE); + if (rv) + return rv; + + kbuf = memdup_user_nul(buf, count); + if (IS_ERR(kbuf)) + return PTR_ERR(kbuf); + + tmp = strchr(kbuf, ' '); + if (!tmp) { + rv = -EINVAL; + goto out; + } + *tmp = '\0'; + tmp++; + + if (!strcmp(kbuf, "add_node")) + rv = do_add_node(tmp, count - (tmp - kbuf)); + else if (!strcmp(kbuf, "remove_node")) + rv = do_remove_node(tmp); + else if (!strcmp(kbuf, "add_property")) + rv = do_add_property(tmp, count - (tmp - kbuf)); + else if (!strcmp(kbuf, "remove_property")) + rv = do_remove_property(tmp, count - (tmp - kbuf)); + else if (!strcmp(kbuf, "update_property")) + rv = do_update_property(tmp, count - (tmp - kbuf)); + else + rv = -EINVAL; +out: + kfree(kbuf); + return rv ? rv : count; +} + +static const struct proc_ops ofdt_proc_ops = { + .proc_write = ofdt_write, + .proc_lseek = noop_llseek, +}; + +/* create /proc/powerpc/ofdt write-only by root */ +static int proc_ppc64_create_ofdt(void) +{ + struct proc_dir_entry *ent; + + ent = proc_create("powerpc/ofdt", 0200, NULL, &ofdt_proc_ops); + if (ent) + proc_set_size(ent, 0); + + return 0; +} +machine_device_initcall(pseries, proc_ppc64_create_ofdt); diff --git a/arch/powerpc/platforms/pseries/rng.c b/arch/powerpc/platforms/pseries/rng.c new file mode 100644 index 000000000..6ddfdeaac --- /dev/null +++ b/arch/powerpc/platforms/pseries/rng.c @@ -0,0 +1,37 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright 2013, Michael Ellerman, IBM Corporation. + */ + +#define pr_fmt(fmt) "pseries-rng: " fmt + +#include <linux/kernel.h> +#include <linux/of.h> +#include <asm/archrandom.h> +#include <asm/machdep.h> +#include <asm/plpar_wrappers.h> +#include "pseries.h" + + +static int pseries_get_random_long(unsigned long *v) +{ + unsigned long retbuf[PLPAR_HCALL_BUFSIZE]; + + if (plpar_hcall(H_RANDOM, retbuf) == H_SUCCESS) { + *v = retbuf[0]; + return 1; + } + + return 0; +} + +void __init pseries_rng_init(void) +{ + struct device_node *dn; + + dn = of_find_compatible_node(NULL, NULL, "ibm,random"); + if (!dn) + return; + ppc_md.get_random_seed = pseries_get_random_long; + of_node_put(dn); +} diff --git a/arch/powerpc/platforms/pseries/rtas-fadump.c b/arch/powerpc/platforms/pseries/rtas-fadump.c new file mode 100644 index 000000000..b5853e9fc --- /dev/null +++ b/arch/powerpc/platforms/pseries/rtas-fadump.c @@ -0,0 +1,557 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Firmware-Assisted Dump support on POWERVM platform. + * + * Copyright 2011, Mahesh Salgaonkar, IBM Corporation. + * Copyright 2019, Hari Bathini, IBM Corporation. + */ + +#define pr_fmt(fmt) "rtas fadump: " fmt + +#include <linux/string.h> +#include <linux/memblock.h> +#include <linux/delay.h> +#include <linux/seq_file.h> +#include <linux/crash_dump.h> +#include <linux/of.h> +#include <linux/of_fdt.h> + +#include <asm/page.h> +#include <asm/rtas.h> +#include <asm/fadump.h> +#include <asm/fadump-internal.h> + +#include "rtas-fadump.h" + +static struct rtas_fadump_mem_struct fdm; +static const struct rtas_fadump_mem_struct *fdm_active; + +static void rtas_fadump_update_config(struct fw_dump *fadump_conf, + const struct rtas_fadump_mem_struct *fdm) +{ + fadump_conf->boot_mem_dest_addr = + be64_to_cpu(fdm->rmr_region.destination_address); + + fadump_conf->fadumphdr_addr = (fadump_conf->boot_mem_dest_addr + + fadump_conf->boot_memory_size); +} + +/* + * This function is called in the capture kernel to get configuration details + * setup in the first kernel and passed to the f/w. + */ +static void __init rtas_fadump_get_config(struct fw_dump *fadump_conf, + const struct rtas_fadump_mem_struct *fdm) +{ + fadump_conf->boot_mem_addr[0] = + be64_to_cpu(fdm->rmr_region.source_address); + fadump_conf->boot_mem_sz[0] = be64_to_cpu(fdm->rmr_region.source_len); + fadump_conf->boot_memory_size = fadump_conf->boot_mem_sz[0]; + + fadump_conf->boot_mem_top = fadump_conf->boot_memory_size; + fadump_conf->boot_mem_regs_cnt = 1; + + /* + * Start address of reserve dump area (permanent reservation) for + * re-registering FADump after dump capture. + */ + fadump_conf->reserve_dump_area_start = + be64_to_cpu(fdm->cpu_state_data.destination_address); + + rtas_fadump_update_config(fadump_conf, fdm); +} + +static u64 rtas_fadump_init_mem_struct(struct fw_dump *fadump_conf) +{ + u64 addr = fadump_conf->reserve_dump_area_start; + + memset(&fdm, 0, sizeof(struct rtas_fadump_mem_struct)); + addr = addr & PAGE_MASK; + + fdm.header.dump_format_version = cpu_to_be32(0x00000001); + fdm.header.dump_num_sections = cpu_to_be16(3); + fdm.header.dump_status_flag = 0; + fdm.header.offset_first_dump_section = + cpu_to_be32((u32)offsetof(struct rtas_fadump_mem_struct, + cpu_state_data)); + + /* + * Fields for disk dump option. + * We are not using disk dump option, hence set these fields to 0. + */ + fdm.header.dd_block_size = 0; + fdm.header.dd_block_offset = 0; + fdm.header.dd_num_blocks = 0; + fdm.header.dd_offset_disk_path = 0; + + /* set 0 to disable an automatic dump-reboot. */ + fdm.header.max_time_auto = 0; + + /* Kernel dump sections */ + /* cpu state data section. */ + fdm.cpu_state_data.request_flag = + cpu_to_be32(RTAS_FADUMP_REQUEST_FLAG); + fdm.cpu_state_data.source_data_type = + cpu_to_be16(RTAS_FADUMP_CPU_STATE_DATA); + fdm.cpu_state_data.source_address = 0; + fdm.cpu_state_data.source_len = + cpu_to_be64(fadump_conf->cpu_state_data_size); + fdm.cpu_state_data.destination_address = cpu_to_be64(addr); + addr += fadump_conf->cpu_state_data_size; + + /* hpte region section */ + fdm.hpte_region.request_flag = cpu_to_be32(RTAS_FADUMP_REQUEST_FLAG); + fdm.hpte_region.source_data_type = + cpu_to_be16(RTAS_FADUMP_HPTE_REGION); + fdm.hpte_region.source_address = 0; + fdm.hpte_region.source_len = + cpu_to_be64(fadump_conf->hpte_region_size); + fdm.hpte_region.destination_address = cpu_to_be64(addr); + addr += fadump_conf->hpte_region_size; + + /* + * Align boot memory area destination address to page boundary to + * be able to mmap read this area in the vmcore. + */ + addr = PAGE_ALIGN(addr); + + /* RMA region section */ + fdm.rmr_region.request_flag = cpu_to_be32(RTAS_FADUMP_REQUEST_FLAG); + fdm.rmr_region.source_data_type = + cpu_to_be16(RTAS_FADUMP_REAL_MODE_REGION); + fdm.rmr_region.source_address = cpu_to_be64(0); + fdm.rmr_region.source_len = cpu_to_be64(fadump_conf->boot_memory_size); + fdm.rmr_region.destination_address = cpu_to_be64(addr); + addr += fadump_conf->boot_memory_size; + + rtas_fadump_update_config(fadump_conf, &fdm); + + return addr; +} + +static u64 rtas_fadump_get_bootmem_min(void) +{ + return RTAS_FADUMP_MIN_BOOT_MEM; +} + +static int rtas_fadump_register(struct fw_dump *fadump_conf) +{ + unsigned int wait_time; + int rc, err = -EIO; + + /* TODO: Add upper time limit for the delay */ + do { + rc = rtas_call(fadump_conf->ibm_configure_kernel_dump, 3, 1, + NULL, FADUMP_REGISTER, &fdm, + sizeof(struct rtas_fadump_mem_struct)); + + wait_time = rtas_busy_delay_time(rc); + if (wait_time) + mdelay(wait_time); + + } while (wait_time); + + switch (rc) { + case 0: + pr_info("Registration is successful!\n"); + fadump_conf->dump_registered = 1; + err = 0; + break; + case -1: + pr_err("Failed to register. Hardware Error(%d).\n", rc); + break; + case -3: + if (!is_fadump_boot_mem_contiguous()) + pr_err("Can't have holes in boot memory area.\n"); + else if (!is_fadump_reserved_mem_contiguous()) + pr_err("Can't have holes in reserved memory area.\n"); + + pr_err("Failed to register. Parameter Error(%d).\n", rc); + err = -EINVAL; + break; + case -9: + pr_err("Already registered!\n"); + fadump_conf->dump_registered = 1; + err = -EEXIST; + break; + default: + pr_err("Failed to register. Unknown Error(%d).\n", rc); + break; + } + + return err; +} + +static int rtas_fadump_unregister(struct fw_dump *fadump_conf) +{ + unsigned int wait_time; + int rc; + + /* TODO: Add upper time limit for the delay */ + do { + rc = rtas_call(fadump_conf->ibm_configure_kernel_dump, 3, 1, + NULL, FADUMP_UNREGISTER, &fdm, + sizeof(struct rtas_fadump_mem_struct)); + + wait_time = rtas_busy_delay_time(rc); + if (wait_time) + mdelay(wait_time); + } while (wait_time); + + if (rc) { + pr_err("Failed to un-register - unexpected error(%d).\n", rc); + return -EIO; + } + + fadump_conf->dump_registered = 0; + return 0; +} + +static int rtas_fadump_invalidate(struct fw_dump *fadump_conf) +{ + unsigned int wait_time; + int rc; + + /* TODO: Add upper time limit for the delay */ + do { + rc = rtas_call(fadump_conf->ibm_configure_kernel_dump, 3, 1, + NULL, FADUMP_INVALIDATE, fdm_active, + sizeof(struct rtas_fadump_mem_struct)); + + wait_time = rtas_busy_delay_time(rc); + if (wait_time) + mdelay(wait_time); + } while (wait_time); + + if (rc) { + pr_err("Failed to invalidate - unexpected error (%d).\n", rc); + return -EIO; + } + + fadump_conf->dump_active = 0; + fdm_active = NULL; + return 0; +} + +#define RTAS_FADUMP_GPR_MASK 0xffffff0000000000 +static inline int rtas_fadump_gpr_index(u64 id) +{ + char str[3]; + int i = -1; + + if ((id & RTAS_FADUMP_GPR_MASK) == fadump_str_to_u64("GPR")) { + /* get the digits at the end */ + id &= ~RTAS_FADUMP_GPR_MASK; + id >>= 24; + str[2] = '\0'; + str[1] = id & 0xff; + str[0] = (id >> 8) & 0xff; + if (kstrtoint(str, 10, &i)) + i = -EINVAL; + if (i > 31) + i = -1; + } + return i; +} + +static void __init rtas_fadump_set_regval(struct pt_regs *regs, u64 reg_id, u64 reg_val) +{ + int i; + + i = rtas_fadump_gpr_index(reg_id); + if (i >= 0) + regs->gpr[i] = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("NIA")) + regs->nip = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("MSR")) + regs->msr = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("CTR")) + regs->ctr = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("LR")) + regs->link = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("XER")) + regs->xer = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("CR")) + regs->ccr = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("DAR")) + regs->dar = (unsigned long)reg_val; + else if (reg_id == fadump_str_to_u64("DSISR")) + regs->dsisr = (unsigned long)reg_val; +} + +static struct rtas_fadump_reg_entry* __init +rtas_fadump_read_regs(struct rtas_fadump_reg_entry *reg_entry, + struct pt_regs *regs) +{ + memset(regs, 0, sizeof(struct pt_regs)); + + while (be64_to_cpu(reg_entry->reg_id) != fadump_str_to_u64("CPUEND")) { + rtas_fadump_set_regval(regs, be64_to_cpu(reg_entry->reg_id), + be64_to_cpu(reg_entry->reg_value)); + reg_entry++; + } + reg_entry++; + return reg_entry; +} + +/* + * Read CPU state dump data and convert it into ELF notes. + * The CPU dump starts with magic number "REGSAVE". NumCpusOffset should be + * used to access the data to allow for additional fields to be added without + * affecting compatibility. Each list of registers for a CPU starts with + * "CPUSTRT" and ends with "CPUEND". Each register entry is of 16 bytes, + * 8 Byte ASCII identifier and 8 Byte register value. The register entry + * with identifier "CPUSTRT" and "CPUEND" contains 4 byte cpu id as part + * of register value. For more details refer to PAPR document. + * + * Only for the crashing cpu we ignore the CPU dump data and get exact + * state from fadump crash info structure populated by first kernel at the + * time of crash. + */ +static int __init rtas_fadump_build_cpu_notes(struct fw_dump *fadump_conf) +{ + struct rtas_fadump_reg_save_area_header *reg_header; + struct fadump_crash_info_header *fdh = NULL; + struct rtas_fadump_reg_entry *reg_entry; + u32 num_cpus, *note_buf; + int i, rc = 0, cpu = 0; + struct pt_regs regs; + unsigned long addr; + void *vaddr; + + addr = be64_to_cpu(fdm_active->cpu_state_data.destination_address); + vaddr = __va(addr); + + reg_header = vaddr; + if (be64_to_cpu(reg_header->magic_number) != + fadump_str_to_u64("REGSAVE")) { + pr_err("Unable to read register save area.\n"); + return -ENOENT; + } + + pr_debug("--------CPU State Data------------\n"); + pr_debug("Magic Number: %llx\n", be64_to_cpu(reg_header->magic_number)); + pr_debug("NumCpuOffset: %x\n", be32_to_cpu(reg_header->num_cpu_offset)); + + vaddr += be32_to_cpu(reg_header->num_cpu_offset); + num_cpus = be32_to_cpu(*((__be32 *)(vaddr))); + pr_debug("NumCpus : %u\n", num_cpus); + vaddr += sizeof(u32); + reg_entry = (struct rtas_fadump_reg_entry *)vaddr; + + rc = fadump_setup_cpu_notes_buf(num_cpus); + if (rc != 0) + return rc; + + note_buf = (u32 *)fadump_conf->cpu_notes_buf_vaddr; + + if (fadump_conf->fadumphdr_addr) + fdh = __va(fadump_conf->fadumphdr_addr); + + for (i = 0; i < num_cpus; i++) { + if (be64_to_cpu(reg_entry->reg_id) != + fadump_str_to_u64("CPUSTRT")) { + pr_err("Unable to read CPU state data\n"); + rc = -ENOENT; + goto error_out; + } + /* Lower 4 bytes of reg_value contains logical cpu id */ + cpu = (be64_to_cpu(reg_entry->reg_value) & + RTAS_FADUMP_CPU_ID_MASK); + if (fdh && !cpumask_test_cpu(cpu, &fdh->cpu_mask)) { + RTAS_FADUMP_SKIP_TO_NEXT_CPU(reg_entry); + continue; + } + pr_debug("Reading register data for cpu %d...\n", cpu); + if (fdh && fdh->crashing_cpu == cpu) { + regs = fdh->regs; + note_buf = fadump_regs_to_elf_notes(note_buf, ®s); + RTAS_FADUMP_SKIP_TO_NEXT_CPU(reg_entry); + } else { + reg_entry++; + reg_entry = rtas_fadump_read_regs(reg_entry, ®s); + note_buf = fadump_regs_to_elf_notes(note_buf, ®s); + } + } + final_note(note_buf); + + if (fdh) { + pr_debug("Updating elfcore header (%llx) with cpu notes\n", + fdh->elfcorehdr_addr); + fadump_update_elfcore_header(__va(fdh->elfcorehdr_addr)); + } + return 0; + +error_out: + fadump_free_cpu_notes_buf(); + return rc; + +} + +/* + * Validate and process the dump data stored by firmware before exporting + * it through '/proc/vmcore'. + */ +static int __init rtas_fadump_process(struct fw_dump *fadump_conf) +{ + struct fadump_crash_info_header *fdh; + int rc = 0; + + if (!fdm_active || !fadump_conf->fadumphdr_addr) + return -EINVAL; + + /* Check if the dump data is valid. */ + if ((be16_to_cpu(fdm_active->header.dump_status_flag) == + RTAS_FADUMP_ERROR_FLAG) || + (fdm_active->cpu_state_data.error_flags != 0) || + (fdm_active->rmr_region.error_flags != 0)) { + pr_err("Dump taken by platform is not valid\n"); + return -EINVAL; + } + if ((fdm_active->rmr_region.bytes_dumped != + fdm_active->rmr_region.source_len) || + !fdm_active->cpu_state_data.bytes_dumped) { + pr_err("Dump taken by platform is incomplete\n"); + return -EINVAL; + } + + /* Validate the fadump crash info header */ + fdh = __va(fadump_conf->fadumphdr_addr); + if (fdh->magic_number != FADUMP_CRASH_INFO_MAGIC) { + pr_err("Crash info header is not valid.\n"); + return -EINVAL; + } + + rc = rtas_fadump_build_cpu_notes(fadump_conf); + if (rc) + return rc; + + /* + * We are done validating dump info and elfcore header is now ready + * to be exported. set elfcorehdr_addr so that vmcore module will + * export the elfcore header through '/proc/vmcore'. + */ + elfcorehdr_addr = fdh->elfcorehdr_addr; + + return 0; +} + +static void rtas_fadump_region_show(struct fw_dump *fadump_conf, + struct seq_file *m) +{ + const struct rtas_fadump_section *cpu_data_section; + const struct rtas_fadump_mem_struct *fdm_ptr; + + if (fdm_active) + fdm_ptr = fdm_active; + else + fdm_ptr = &fdm; + + cpu_data_section = &(fdm_ptr->cpu_state_data); + seq_printf(m, "CPU :[%#016llx-%#016llx] %#llx bytes, Dumped: %#llx\n", + be64_to_cpu(cpu_data_section->destination_address), + be64_to_cpu(cpu_data_section->destination_address) + + be64_to_cpu(cpu_data_section->source_len) - 1, + be64_to_cpu(cpu_data_section->source_len), + be64_to_cpu(cpu_data_section->bytes_dumped)); + + seq_printf(m, "HPTE:[%#016llx-%#016llx] %#llx bytes, Dumped: %#llx\n", + be64_to_cpu(fdm_ptr->hpte_region.destination_address), + be64_to_cpu(fdm_ptr->hpte_region.destination_address) + + be64_to_cpu(fdm_ptr->hpte_region.source_len) - 1, + be64_to_cpu(fdm_ptr->hpte_region.source_len), + be64_to_cpu(fdm_ptr->hpte_region.bytes_dumped)); + + seq_printf(m, "DUMP: Src: %#016llx, Dest: %#016llx, ", + be64_to_cpu(fdm_ptr->rmr_region.source_address), + be64_to_cpu(fdm_ptr->rmr_region.destination_address)); + seq_printf(m, "Size: %#llx, Dumped: %#llx bytes\n", + be64_to_cpu(fdm_ptr->rmr_region.source_len), + be64_to_cpu(fdm_ptr->rmr_region.bytes_dumped)); + + /* Dump is active. Show preserved area start address. */ + if (fdm_active) { + seq_printf(m, "\nMemory above %#016llx is reserved for saving crash dump\n", + fadump_conf->boot_mem_top); + } +} + +static void rtas_fadump_trigger(struct fadump_crash_info_header *fdh, + const char *msg) +{ + /* Call ibm,os-term rtas call to trigger firmware assisted dump */ + rtas_os_term((char *)msg); +} + +static struct fadump_ops rtas_fadump_ops = { + .fadump_init_mem_struct = rtas_fadump_init_mem_struct, + .fadump_get_bootmem_min = rtas_fadump_get_bootmem_min, + .fadump_register = rtas_fadump_register, + .fadump_unregister = rtas_fadump_unregister, + .fadump_invalidate = rtas_fadump_invalidate, + .fadump_process = rtas_fadump_process, + .fadump_region_show = rtas_fadump_region_show, + .fadump_trigger = rtas_fadump_trigger, +}; + +void __init rtas_fadump_dt_scan(struct fw_dump *fadump_conf, u64 node) +{ + int i, size, num_sections; + const __be32 *sections; + const __be32 *token; + + /* + * Check if Firmware Assisted dump is supported. if yes, check + * if dump has been initiated on last reboot. + */ + token = of_get_flat_dt_prop(node, "ibm,configure-kernel-dump", NULL); + if (!token) + return; + + fadump_conf->ibm_configure_kernel_dump = be32_to_cpu(*token); + fadump_conf->ops = &rtas_fadump_ops; + fadump_conf->fadump_supported = 1; + + /* Firmware supports 64-bit value for size, align it to pagesize. */ + fadump_conf->max_copy_size = ALIGN_DOWN(U64_MAX, PAGE_SIZE); + + /* + * The 'ibm,kernel-dump' rtas node is present only if there is + * dump data waiting for us. + */ + fdm_active = of_get_flat_dt_prop(node, "ibm,kernel-dump", NULL); + if (fdm_active) { + pr_info("Firmware-assisted dump is active.\n"); + fadump_conf->dump_active = 1; + rtas_fadump_get_config(fadump_conf, (void *)__pa(fdm_active)); + } + + /* Get the sizes required to store dump data for the firmware provided + * dump sections. + * For each dump section type supported, a 32bit cell which defines + * the ID of a supported section followed by two 32 bit cells which + * gives the size of the section in bytes. + */ + sections = of_get_flat_dt_prop(node, "ibm,configure-kernel-dump-sizes", + &size); + + if (!sections) + return; + + num_sections = size / (3 * sizeof(u32)); + + for (i = 0; i < num_sections; i++, sections += 3) { + u32 type = (u32)of_read_number(sections, 1); + + switch (type) { + case RTAS_FADUMP_CPU_STATE_DATA: + fadump_conf->cpu_state_data_size = + of_read_ulong(§ions[1], 2); + break; + case RTAS_FADUMP_HPTE_REGION: + fadump_conf->hpte_region_size = + of_read_ulong(§ions[1], 2); + break; + } + } +} diff --git a/arch/powerpc/platforms/pseries/rtas-fadump.h b/arch/powerpc/platforms/pseries/rtas-fadump.h new file mode 100644 index 000000000..fd59bd7ca --- /dev/null +++ b/arch/powerpc/platforms/pseries/rtas-fadump.h @@ -0,0 +1,114 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Firmware-Assisted Dump support on POWERVM platform. + * + * Copyright 2011, Mahesh Salgaonkar, IBM Corporation. + * Copyright 2019, Hari Bathini, IBM Corporation. + */ + +#ifndef _PSERIES_RTAS_FADUMP_H +#define _PSERIES_RTAS_FADUMP_H + +/* + * On some Power systems where RMO is 128MB, it still requires minimum of + * 256MB for kernel to boot successfully. When kdump infrastructure is + * configured to save vmcore over network, we run into OOM issue while + * loading modules related to network setup. Hence we need additional 64M + * of memory to avoid OOM issue. + */ +#define RTAS_FADUMP_MIN_BOOT_MEM ((0x1UL << 28) + (0x1UL << 26)) + +/* Firmware provided dump sections */ +#define RTAS_FADUMP_CPU_STATE_DATA 0x0001 +#define RTAS_FADUMP_HPTE_REGION 0x0002 +#define RTAS_FADUMP_REAL_MODE_REGION 0x0011 + +/* Dump request flag */ +#define RTAS_FADUMP_REQUEST_FLAG 0x00000001 + +/* Dump status flag */ +#define RTAS_FADUMP_ERROR_FLAG 0x2000 + +/* Kernel Dump section info */ +struct rtas_fadump_section { + __be32 request_flag; + __be16 source_data_type; + __be16 error_flags; + __be64 source_address; + __be64 source_len; + __be64 bytes_dumped; + __be64 destination_address; +}; + +/* ibm,configure-kernel-dump header. */ +struct rtas_fadump_section_header { + __be32 dump_format_version; + __be16 dump_num_sections; + __be16 dump_status_flag; + __be32 offset_first_dump_section; + + /* Fields for disk dump option. */ + __be32 dd_block_size; + __be64 dd_block_offset; + __be64 dd_num_blocks; + __be32 dd_offset_disk_path; + + /* Maximum time allowed to prevent an automatic dump-reboot. */ + __be32 max_time_auto; +}; + +/* + * Firmware Assisted dump memory structure. This structure is required for + * registering future kernel dump with power firmware through rtas call. + * + * No disk dump option. Hence disk dump path string section is not included. + */ +struct rtas_fadump_mem_struct { + struct rtas_fadump_section_header header; + + /* Kernel dump sections */ + struct rtas_fadump_section cpu_state_data; + struct rtas_fadump_section hpte_region; + + /* + * TODO: Extend multiple boot memory regions support in the kernel + * for this platform. + */ + struct rtas_fadump_section rmr_region; +}; + +/* + * The firmware-assisted dump format. + * + * The register save area is an area in the partition's memory used to preserve + * the register contents (CPU state data) for the active CPUs during a firmware + * assisted dump. The dump format contains register save area header followed + * by register entries. Each list of registers for a CPU starts with "CPUSTRT" + * and ends with "CPUEND". + */ + +/* Register save area header. */ +struct rtas_fadump_reg_save_area_header { + __be64 magic_number; + __be32 version; + __be32 num_cpu_offset; +}; + +/* Register entry. */ +struct rtas_fadump_reg_entry { + __be64 reg_id; + __be64 reg_value; +}; + +/* Utility macros */ +#define RTAS_FADUMP_SKIP_TO_NEXT_CPU(reg_entry) \ +({ \ + while (be64_to_cpu(reg_entry->reg_id) != \ + fadump_str_to_u64("CPUEND")) \ + reg_entry++; \ + reg_entry++; \ +}) + +#define RTAS_FADUMP_CPU_ID_MASK ((1UL << 32) - 1) + +#endif /* _PSERIES_RTAS_FADUMP_H */ diff --git a/arch/powerpc/platforms/pseries/rtas-work-area.c b/arch/powerpc/platforms/pseries/rtas-work-area.c new file mode 100644 index 000000000..b37d52f40 --- /dev/null +++ b/arch/powerpc/platforms/pseries/rtas-work-area.c @@ -0,0 +1,209 @@ +// SPDX-License-Identifier: GPL-2.0-only + +#define pr_fmt(fmt) "rtas-work-area: " fmt + +#include <linux/genalloc.h> +#include <linux/log2.h> +#include <linux/kernel.h> +#include <linux/memblock.h> +#include <linux/mempool.h> +#include <linux/minmax.h> +#include <linux/mutex.h> +#include <linux/numa.h> +#include <linux/sizes.h> +#include <linux/wait.h> + +#include <asm/machdep.h> +#include <asm/rtas-work-area.h> +#include <asm/rtas.h> + +enum { + /* + * Ensure the pool is page-aligned. + */ + RTAS_WORK_AREA_ARENA_ALIGN = PAGE_SIZE, + /* + * Don't let a single allocation claim the whole arena. + */ + RTAS_WORK_AREA_ARENA_SZ = RTAS_WORK_AREA_MAX_ALLOC_SZ * 2, + /* + * The smallest known work area size is for ibm,get-vpd's + * location code argument, which is limited to 79 characters + * plus 1 nul terminator. + * + * PAPR+ 7.3.20 ibm,get-vpd RTAS Call + * PAPR+ 12.3.2.4 Converged Location Code Rules - Length Restrictions + */ + RTAS_WORK_AREA_MIN_ALLOC_SZ = roundup_pow_of_two(80), +}; + +static struct { + struct gen_pool *gen_pool; + char *arena; + struct mutex mutex; /* serializes allocations */ + struct wait_queue_head wqh; + mempool_t descriptor_pool; + bool available; +} rwa_state = { + .mutex = __MUTEX_INITIALIZER(rwa_state.mutex), + .wqh = __WAIT_QUEUE_HEAD_INITIALIZER(rwa_state.wqh), +}; + +/* + * A single work area buffer and descriptor to serve requests early in + * boot before the allocator is fully initialized. We know 4KB is the + * most any boot time user needs (they all call ibm,get-system-parameter). + */ +static bool early_work_area_in_use __initdata; +static char early_work_area_buf[SZ_4K] __initdata __aligned(SZ_4K); +static struct rtas_work_area early_work_area __initdata = { + .buf = early_work_area_buf, + .size = sizeof(early_work_area_buf), +}; + + +static struct rtas_work_area * __init rtas_work_area_alloc_early(size_t size) +{ + WARN_ON(size > early_work_area.size); + WARN_ON(early_work_area_in_use); + early_work_area_in_use = true; + memset(early_work_area.buf, 0, early_work_area.size); + return &early_work_area; +} + +static void __init rtas_work_area_free_early(struct rtas_work_area *work_area) +{ + WARN_ON(work_area != &early_work_area); + WARN_ON(!early_work_area_in_use); + early_work_area_in_use = false; +} + +struct rtas_work_area * __ref __rtas_work_area_alloc(size_t size) +{ + struct rtas_work_area *area; + unsigned long addr; + + might_sleep(); + + /* + * The rtas_work_area_alloc() wrapper enforces this at build + * time. Requests that exceed the arena size will block + * indefinitely. + */ + WARN_ON(size > RTAS_WORK_AREA_MAX_ALLOC_SZ); + + if (!rwa_state.available) + return rtas_work_area_alloc_early(size); + /* + * To ensure FCFS behavior and prevent a high rate of smaller + * requests from starving larger ones, use the mutex to queue + * allocations. + */ + mutex_lock(&rwa_state.mutex); + wait_event(rwa_state.wqh, + (addr = gen_pool_alloc(rwa_state.gen_pool, size)) != 0); + mutex_unlock(&rwa_state.mutex); + + area = mempool_alloc(&rwa_state.descriptor_pool, GFP_KERNEL); + area->buf = (char *)addr; + area->size = size; + + return area; +} + +void __ref rtas_work_area_free(struct rtas_work_area *area) +{ + if (!rwa_state.available) { + rtas_work_area_free_early(area); + return; + } + + gen_pool_free(rwa_state.gen_pool, (unsigned long)area->buf, area->size); + mempool_free(area, &rwa_state.descriptor_pool); + wake_up(&rwa_state.wqh); +} + +/* + * Initialization of the work area allocator happens in two parts. To + * reliably reserve an arena that satisfies RTAS addressing + * requirements, we must perform a memblock allocation early, + * immmediately after RTAS instantiation. Then we have to wait until + * the slab allocator is up before setting up the descriptor mempool + * and adding the arena to a gen_pool. + */ +static __init int rtas_work_area_allocator_init(void) +{ + const unsigned int order = ilog2(RTAS_WORK_AREA_MIN_ALLOC_SZ); + const phys_addr_t pa_start = __pa(rwa_state.arena); + const phys_addr_t pa_end = pa_start + RTAS_WORK_AREA_ARENA_SZ - 1; + struct gen_pool *pool; + const int nid = NUMA_NO_NODE; + int err; + + err = -ENOMEM; + if (!rwa_state.arena) + goto err_out; + + pool = gen_pool_create(order, nid); + if (!pool) + goto err_out; + /* + * All RTAS functions that consume work areas are OK with + * natural alignment, when they have alignment requirements at + * all. + */ + gen_pool_set_algo(pool, gen_pool_first_fit_order_align, NULL); + + err = gen_pool_add(pool, (unsigned long)rwa_state.arena, + RTAS_WORK_AREA_ARENA_SZ, nid); + if (err) + goto err_destroy; + + err = mempool_init_kmalloc_pool(&rwa_state.descriptor_pool, 1, + sizeof(struct rtas_work_area)); + if (err) + goto err_destroy; + + rwa_state.gen_pool = pool; + rwa_state.available = true; + + pr_debug("arena [%pa-%pa] (%uK), min/max alloc sizes %u/%u\n", + &pa_start, &pa_end, + RTAS_WORK_AREA_ARENA_SZ / SZ_1K, + RTAS_WORK_AREA_MIN_ALLOC_SZ, + RTAS_WORK_AREA_MAX_ALLOC_SZ); + + return 0; + +err_destroy: + gen_pool_destroy(pool); +err_out: + return err; +} +machine_arch_initcall(pseries, rtas_work_area_allocator_init); + +/** + * rtas_work_area_reserve_arena() - Reserve memory suitable for RTAS work areas. + */ +void __init rtas_work_area_reserve_arena(const phys_addr_t limit) +{ + const phys_addr_t align = RTAS_WORK_AREA_ARENA_ALIGN; + const phys_addr_t size = RTAS_WORK_AREA_ARENA_SZ; + const phys_addr_t min = MEMBLOCK_LOW_LIMIT; + const int nid = NUMA_NO_NODE; + + /* + * Too early for a machine_is(pseries) check. But PAPR + * effectively mandates that ibm,get-system-parameter is + * present: + * + * R1–7.3.16–1. All platforms must support the System + * Parameters option. + * + * So set up the arena if we find that, with a fallback to + * ibm,configure-connector, just in case. + */ + if (rtas_function_implemented(RTAS_FN_IBM_GET_SYSTEM_PARAMETER) || + rtas_function_implemented(RTAS_FN_IBM_CONFIGURE_CONNECTOR)) + rwa_state.arena = memblock_alloc_try_nid(size, align, min, limit, nid); +} diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c new file mode 100644 index 000000000..ecea85c74 --- /dev/null +++ b/arch/powerpc/platforms/pseries/setup.c @@ -0,0 +1,1162 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * 64-bit pSeries and RS/6000 setup code. + * + * Copyright (C) 1995 Linus Torvalds + * Adapted from 'alpha' version by Gary Thomas + * Modified by Cort Dougan (cort@cs.nmt.edu) + * Modified by PPC64 Team, IBM Corp + */ + +/* + * bootup setup stuff.. + */ + +#include <linux/cpu.h> +#include <linux/errno.h> +#include <linux/platform_device.h> +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/mm.h> +#include <linux/stddef.h> +#include <linux/unistd.h> +#include <linux/user.h> +#include <linux/tty.h> +#include <linux/major.h> +#include <linux/interrupt.h> +#include <linux/reboot.h> +#include <linux/init.h> +#include <linux/ioport.h> +#include <linux/console.h> +#include <linux/pci.h> +#include <linux/utsname.h> +#include <linux/adb.h> +#include <linux/export.h> +#include <linux/delay.h> +#include <linux/irq.h> +#include <linux/seq_file.h> +#include <linux/root_dev.h> +#include <linux/of.h> +#include <linux/of_irq.h> +#include <linux/of_pci.h> +#include <linux/memblock.h> +#include <linux/swiotlb.h> +#include <linux/seq_buf.h> + +#include <asm/mmu.h> +#include <asm/processor.h> +#include <asm/io.h> +#include <asm/rtas.h> +#include <asm/pci-bridge.h> +#include <asm/iommu.h> +#include <asm/dma.h> +#include <asm/machdep.h> +#include <asm/irq.h> +#include <asm/time.h> +#include <asm/nvram.h> +#include <asm/pmc.h> +#include <asm/xics.h> +#include <asm/xive.h> +#include <asm/papr-sysparm.h> +#include <asm/ppc-pci.h> +#include <asm/i8259.h> +#include <asm/udbg.h> +#include <asm/smp.h> +#include <asm/firmware.h> +#include <asm/eeh.h> +#include <asm/reg.h> +#include <asm/plpar_wrappers.h> +#include <asm/kexec.h> +#include <asm/isa-bridge.h> +#include <asm/security_features.h> +#include <asm/asm-const.h> +#include <asm/idle.h> +#include <asm/swiotlb.h> +#include <asm/svm.h> +#include <asm/dtl.h> +#include <asm/hvconsole.h> +#include <asm/setup.h> + +#include "pseries.h" + +DEFINE_STATIC_KEY_FALSE(shared_processor); +EXPORT_SYMBOL(shared_processor); + +#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING +struct static_key paravirt_steal_enabled; +struct static_key paravirt_steal_rq_enabled; + +static bool steal_acc = true; +static int __init parse_no_stealacc(char *arg) +{ + steal_acc = false; + return 0; +} + +early_param("no-steal-acc", parse_no_stealacc); +#endif + +int CMO_PrPSP = -1; +int CMO_SecPSP = -1; +unsigned long CMO_PageSize = (ASM_CONST(1) << IOMMU_PAGE_SHIFT_4K); +EXPORT_SYMBOL(CMO_PageSize); + +int fwnmi_active; /* TRUE if an FWNMI handler is present */ +int ibm_nmi_interlock_token; +u32 pseries_security_flavor; + +static void pSeries_show_cpuinfo(struct seq_file *m) +{ + struct device_node *root; + const char *model = ""; + + root = of_find_node_by_path("/"); + if (root) + model = of_get_property(root, "model", NULL); + seq_printf(m, "machine\t\t: CHRP %s\n", model); + of_node_put(root); + if (radix_enabled()) + seq_printf(m, "MMU\t\t: Radix\n"); + else + seq_printf(m, "MMU\t\t: Hash\n"); +} + +/* Initialize firmware assisted non-maskable interrupts if + * the firmware supports this feature. + */ +static void __init fwnmi_init(void) +{ + unsigned long system_reset_addr, machine_check_addr; + u8 *mce_data_buf; + unsigned int i; + int nr_cpus = num_possible_cpus(); +#ifdef CONFIG_PPC_64S_HASH_MMU + struct slb_entry *slb_ptr; + size_t size; +#endif + int ibm_nmi_register_token; + + ibm_nmi_register_token = rtas_function_token(RTAS_FN_IBM_NMI_REGISTER); + if (ibm_nmi_register_token == RTAS_UNKNOWN_SERVICE) + return; + + ibm_nmi_interlock_token = rtas_function_token(RTAS_FN_IBM_NMI_INTERLOCK); + if (WARN_ON(ibm_nmi_interlock_token == RTAS_UNKNOWN_SERVICE)) + return; + + /* If the kernel's not linked at zero we point the firmware at low + * addresses anyway, and use a trampoline to get to the real code. */ + system_reset_addr = __pa(system_reset_fwnmi) - PHYSICAL_START; + machine_check_addr = __pa(machine_check_fwnmi) - PHYSICAL_START; + + if (0 == rtas_call(ibm_nmi_register_token, 2, 1, NULL, + system_reset_addr, machine_check_addr)) + fwnmi_active = 1; + + /* + * Allocate a chunk for per cpu buffer to hold rtas errorlog. + * It will be used in real mode mce handler, hence it needs to be + * below RMA. + */ + mce_data_buf = memblock_alloc_try_nid_raw(RTAS_ERROR_LOG_MAX * nr_cpus, + RTAS_ERROR_LOG_MAX, MEMBLOCK_LOW_LIMIT, + ppc64_rma_size, NUMA_NO_NODE); + if (!mce_data_buf) + panic("Failed to allocate %d bytes below %pa for MCE buffer\n", + RTAS_ERROR_LOG_MAX * nr_cpus, &ppc64_rma_size); + + for_each_possible_cpu(i) { + paca_ptrs[i]->mce_data_buf = mce_data_buf + + (RTAS_ERROR_LOG_MAX * i); + } + +#ifdef CONFIG_PPC_64S_HASH_MMU + if (!radix_enabled()) { + /* Allocate per cpu area to save old slb contents during MCE */ + size = sizeof(struct slb_entry) * mmu_slb_size * nr_cpus; + slb_ptr = memblock_alloc_try_nid_raw(size, + sizeof(struct slb_entry), MEMBLOCK_LOW_LIMIT, + ppc64_rma_size, NUMA_NO_NODE); + if (!slb_ptr) + panic("Failed to allocate %zu bytes below %pa for slb area\n", + size, &ppc64_rma_size); + + for_each_possible_cpu(i) + paca_ptrs[i]->mce_faulty_slbs = slb_ptr + (mmu_slb_size * i); + } +#endif +} + +/* + * Affix a device for the first timer to the platform bus if + * we have firmware support for the H_WATCHDOG hypercall. + */ +static __init int pseries_wdt_init(void) +{ + if (firmware_has_feature(FW_FEATURE_WATCHDOG)) + platform_device_register_simple("pseries-wdt", 0, NULL, 0); + return 0; +} +machine_subsys_initcall(pseries, pseries_wdt_init); + +static void pseries_8259_cascade(struct irq_desc *desc) +{ + struct irq_chip *chip = irq_desc_get_chip(desc); + unsigned int cascade_irq = i8259_irq(); + + if (cascade_irq) + generic_handle_irq(cascade_irq); + + chip->irq_eoi(&desc->irq_data); +} + +static void __init pseries_setup_i8259_cascade(void) +{ + struct device_node *np, *old, *found = NULL; + unsigned int cascade; + const u32 *addrp; + unsigned long intack = 0; + int naddr; + + for_each_node_by_type(np, "interrupt-controller") { + if (of_device_is_compatible(np, "chrp,iic")) { + found = np; + break; + } + } + + if (found == NULL) { + printk(KERN_DEBUG "pic: no ISA interrupt controller\n"); + return; + } + + cascade = irq_of_parse_and_map(found, 0); + if (!cascade) { + printk(KERN_ERR "pic: failed to map cascade interrupt"); + return; + } + pr_debug("pic: cascade mapped to irq %d\n", cascade); + + for (old = of_node_get(found); old != NULL ; old = np) { + np = of_get_parent(old); + of_node_put(old); + if (np == NULL) + break; + if (!of_node_name_eq(np, "pci")) + continue; + addrp = of_get_property(np, "8259-interrupt-acknowledge", NULL); + if (addrp == NULL) + continue; + naddr = of_n_addr_cells(np); + intack = addrp[naddr-1]; + if (naddr > 1) + intack |= ((unsigned long)addrp[naddr-2]) << 32; + } + if (intack) + printk(KERN_DEBUG "pic: PCI 8259 intack at 0x%016lx\n", intack); + i8259_init(found, intack); + of_node_put(found); + irq_set_chained_handler(cascade, pseries_8259_cascade); +} + +static void __init pseries_init_irq(void) +{ + /* Try using a XIVE if available, otherwise use a XICS */ + if (!xive_spapr_init()) { + xics_init(); + pseries_setup_i8259_cascade(); + } +} + +static void pseries_lpar_enable_pmcs(void) +{ + unsigned long set, reset; + + set = 1UL << 63; + reset = 0; + plpar_hcall_norets(H_PERFMON, set, reset); +} + +static int pci_dn_reconfig_notifier(struct notifier_block *nb, unsigned long action, void *data) +{ + struct of_reconfig_data *rd = data; + struct device_node *parent, *np = rd->dn; + struct pci_dn *pdn; + int err = NOTIFY_OK; + + switch (action) { + case OF_RECONFIG_ATTACH_NODE: + parent = of_get_parent(np); + pdn = parent ? PCI_DN(parent) : NULL; + if (pdn) + pci_add_device_node_info(pdn->phb, np); + + of_node_put(parent); + break; + case OF_RECONFIG_DETACH_NODE: + pdn = PCI_DN(np); + if (pdn) + list_del(&pdn->list); + break; + default: + err = NOTIFY_DONE; + break; + } + return err; +} + +static struct notifier_block pci_dn_reconfig_nb = { + .notifier_call = pci_dn_reconfig_notifier, +}; + +struct kmem_cache *dtl_cache; + +#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE +/* + * Allocate space for the dispatch trace log for all possible cpus + * and register the buffers with the hypervisor. This is used for + * computing time stolen by the hypervisor. + */ +static int alloc_dispatch_logs(void) +{ + if (!firmware_has_feature(FW_FEATURE_SPLPAR)) + return 0; + + if (!dtl_cache) + return 0; + + alloc_dtl_buffers(0); + + /* Register the DTL for the current (boot) cpu */ + register_dtl_buffer(smp_processor_id()); + + return 0; +} +#else /* !CONFIG_VIRT_CPU_ACCOUNTING_NATIVE */ +static inline int alloc_dispatch_logs(void) +{ + return 0; +} +#endif /* CONFIG_VIRT_CPU_ACCOUNTING_NATIVE */ + +static int alloc_dispatch_log_kmem_cache(void) +{ + void (*ctor)(void *) = get_dtl_cache_ctor(); + + dtl_cache = kmem_cache_create("dtl", DISPATCH_LOG_BYTES, + DISPATCH_LOG_BYTES, 0, ctor); + if (!dtl_cache) { + pr_warn("Failed to create dispatch trace log buffer cache\n"); + pr_warn("Stolen time statistics will be unreliable\n"); + return 0; + } + + return alloc_dispatch_logs(); +} +machine_early_initcall(pseries, alloc_dispatch_log_kmem_cache); + +DEFINE_PER_CPU(u64, idle_spurr_cycles); +DEFINE_PER_CPU(u64, idle_entry_purr_snap); +DEFINE_PER_CPU(u64, idle_entry_spurr_snap); +static void pseries_lpar_idle(void) +{ + /* + * Default handler to go into low thread priority and possibly + * low power mode by ceding processor to hypervisor + */ + + if (!prep_irq_for_idle()) + return; + + /* Indicate to hypervisor that we are idle. */ + pseries_idle_prolog(); + + /* + * Yield the processor to the hypervisor. We return if + * an external interrupt occurs (which are driven prior + * to returning here) or if a prod occurs from another + * processor. When returning here, external interrupts + * are enabled. + */ + cede_processor(); + + pseries_idle_epilog(); +} + +static bool pseries_reloc_on_exception_enabled; + +bool pseries_reloc_on_exception(void) +{ + return pseries_reloc_on_exception_enabled; +} +EXPORT_SYMBOL_GPL(pseries_reloc_on_exception); + +/* + * Enable relocation on during exceptions. This has partition wide scope and + * may take a while to complete, if it takes longer than one second we will + * just give up rather than wasting any more time on this - if that turns out + * to ever be a problem in practice we can move this into a kernel thread to + * finish off the process later in boot. + */ +bool pseries_enable_reloc_on_exc(void) +{ + long rc; + unsigned int delay, total_delay = 0; + + while (1) { + rc = enable_reloc_on_exceptions(); + if (!H_IS_LONG_BUSY(rc)) { + if (rc == H_P2) { + pr_info("Relocation on exceptions not" + " supported\n"); + return false; + } else if (rc != H_SUCCESS) { + pr_warn("Unable to enable relocation" + " on exceptions: %ld\n", rc); + return false; + } + pseries_reloc_on_exception_enabled = true; + return true; + } + + delay = get_longbusy_msecs(rc); + total_delay += delay; + if (total_delay > 1000) { + pr_warn("Warning: Giving up waiting to enable " + "relocation on exceptions (%u msec)!\n", + total_delay); + return false; + } + + mdelay(delay); + } +} +EXPORT_SYMBOL(pseries_enable_reloc_on_exc); + +void pseries_disable_reloc_on_exc(void) +{ + long rc; + + while (1) { + rc = disable_reloc_on_exceptions(); + if (!H_IS_LONG_BUSY(rc)) + break; + mdelay(get_longbusy_msecs(rc)); + } + if (rc == H_SUCCESS) + pseries_reloc_on_exception_enabled = false; + else + pr_warn("Warning: Failed to disable relocation on exceptions: %ld\n", + rc); +} +EXPORT_SYMBOL(pseries_disable_reloc_on_exc); + +#ifdef __LITTLE_ENDIAN__ +void pseries_big_endian_exceptions(void) +{ + long rc; + + while (1) { + rc = enable_big_endian_exceptions(); + if (!H_IS_LONG_BUSY(rc)) + break; + mdelay(get_longbusy_msecs(rc)); + } + + /* + * At this point it is unlikely panic() will get anything + * out to the user, since this is called very late in kexec + * but at least this will stop us from continuing on further + * and creating an even more difficult to debug situation. + * + * There is a known problem when kdump'ing, if cpus are offline + * the above call will fail. Rather than panicking again, keep + * going and hope the kdump kernel is also little endian, which + * it usually is. + */ + if (rc && !kdump_in_progress()) + panic("Could not enable big endian exceptions"); +} + +void __init pseries_little_endian_exceptions(void) +{ + long rc; + + while (1) { + rc = enable_little_endian_exceptions(); + if (!H_IS_LONG_BUSY(rc)) + break; + mdelay(get_longbusy_msecs(rc)); + } + if (rc) { + ppc_md.progress("H_SET_MODE LE exception fail", 0); + panic("Could not enable little endian exceptions"); + } +} +#endif + +static void __init pSeries_discover_phbs(void) +{ + struct device_node *node; + struct pci_controller *phb; + struct device_node *root = of_find_node_by_path("/"); + + for_each_child_of_node(root, node) { + if (!of_node_is_type(node, "pci") && + !of_node_is_type(node, "pciex")) + continue; + + phb = pcibios_alloc_controller(node); + if (!phb) + continue; + rtas_setup_phb(phb); + pci_process_bridge_OF_ranges(phb, node, 0); + isa_bridge_find_early(phb); + phb->controller_ops = pseries_pci_controller_ops; + + /* create pci_dn's for DT nodes under this PHB */ + pci_devs_phb_init_dynamic(phb); + + pseries_msi_allocate_domains(phb); + } + + of_node_put(root); + + /* + * PCI_PROBE_ONLY and PCI_REASSIGN_ALL_BUS can be set via properties + * in chosen. + */ + of_pci_check_probe_only(); +} + +static void init_cpu_char_feature_flags(struct h_cpu_char_result *result) +{ + /* + * The features below are disabled by default, so we instead look to see + * if firmware has *enabled* them, and set them if so. + */ + if (result->character & H_CPU_CHAR_SPEC_BAR_ORI31) + security_ftr_set(SEC_FTR_SPEC_BAR_ORI31); + + if (result->character & H_CPU_CHAR_BCCTRL_SERIALISED) + security_ftr_set(SEC_FTR_BCCTRL_SERIALISED); + + if (result->character & H_CPU_CHAR_L1D_FLUSH_ORI30) + security_ftr_set(SEC_FTR_L1D_FLUSH_ORI30); + + if (result->character & H_CPU_CHAR_L1D_FLUSH_TRIG2) + security_ftr_set(SEC_FTR_L1D_FLUSH_TRIG2); + + if (result->character & H_CPU_CHAR_L1D_THREAD_PRIV) + security_ftr_set(SEC_FTR_L1D_THREAD_PRIV); + + if (result->character & H_CPU_CHAR_COUNT_CACHE_DISABLED) + security_ftr_set(SEC_FTR_COUNT_CACHE_DISABLED); + + if (result->character & H_CPU_CHAR_BCCTR_FLUSH_ASSIST) + security_ftr_set(SEC_FTR_BCCTR_FLUSH_ASSIST); + + if (result->character & H_CPU_CHAR_BCCTR_LINK_FLUSH_ASSIST) + security_ftr_set(SEC_FTR_BCCTR_LINK_FLUSH_ASSIST); + + if (result->behaviour & H_CPU_BEHAV_FLUSH_COUNT_CACHE) + security_ftr_set(SEC_FTR_FLUSH_COUNT_CACHE); + + if (result->behaviour & H_CPU_BEHAV_FLUSH_LINK_STACK) + security_ftr_set(SEC_FTR_FLUSH_LINK_STACK); + + /* + * The features below are enabled by default, so we instead look to see + * if firmware has *disabled* them, and clear them if so. + * H_CPU_BEHAV_FAVOUR_SECURITY_H could be set only if + * H_CPU_BEHAV_FAVOUR_SECURITY is. + */ + if (!(result->behaviour & H_CPU_BEHAV_FAVOUR_SECURITY)) { + security_ftr_clear(SEC_FTR_FAVOUR_SECURITY); + pseries_security_flavor = 0; + } else if (result->behaviour & H_CPU_BEHAV_FAVOUR_SECURITY_H) + pseries_security_flavor = 1; + else + pseries_security_flavor = 2; + + if (!(result->behaviour & H_CPU_BEHAV_L1D_FLUSH_PR)) + security_ftr_clear(SEC_FTR_L1D_FLUSH_PR); + + if (result->behaviour & H_CPU_BEHAV_NO_L1D_FLUSH_ENTRY) + security_ftr_clear(SEC_FTR_L1D_FLUSH_ENTRY); + + if (result->behaviour & H_CPU_BEHAV_NO_L1D_FLUSH_UACCESS) + security_ftr_clear(SEC_FTR_L1D_FLUSH_UACCESS); + + if (result->behaviour & H_CPU_BEHAV_NO_STF_BARRIER) + security_ftr_clear(SEC_FTR_STF_BARRIER); + + if (!(result->behaviour & H_CPU_BEHAV_BNDS_CHK_SPEC_BAR)) + security_ftr_clear(SEC_FTR_BNDS_CHK_SPEC_BAR); +} + +void pseries_setup_security_mitigations(void) +{ + struct h_cpu_char_result result; + enum l1d_flush_type types; + bool enable; + long rc; + + /* + * Set features to the defaults assumed by init_cpu_char_feature_flags() + * so it can set/clear again any features that might have changed after + * migration, and in case the hypercall fails and it is not even called. + */ + powerpc_security_features = SEC_FTR_DEFAULT; + + rc = plpar_get_cpu_characteristics(&result); + if (rc == H_SUCCESS) + init_cpu_char_feature_flags(&result); + + /* + * We're the guest so this doesn't apply to us, clear it to simplify + * handling of it elsewhere. + */ + security_ftr_clear(SEC_FTR_L1D_FLUSH_HV); + + types = L1D_FLUSH_FALLBACK; + + if (security_ftr_enabled(SEC_FTR_L1D_FLUSH_TRIG2)) + types |= L1D_FLUSH_MTTRIG; + + if (security_ftr_enabled(SEC_FTR_L1D_FLUSH_ORI30)) + types |= L1D_FLUSH_ORI; + + enable = security_ftr_enabled(SEC_FTR_FAVOUR_SECURITY) && \ + security_ftr_enabled(SEC_FTR_L1D_FLUSH_PR); + + setup_rfi_flush(types, enable); + setup_count_cache_flush(); + + enable = security_ftr_enabled(SEC_FTR_FAVOUR_SECURITY) && + security_ftr_enabled(SEC_FTR_L1D_FLUSH_ENTRY); + setup_entry_flush(enable); + + enable = security_ftr_enabled(SEC_FTR_FAVOUR_SECURITY) && + security_ftr_enabled(SEC_FTR_L1D_FLUSH_UACCESS); + setup_uaccess_flush(enable); + + setup_stf_barrier(); +} + +#ifdef CONFIG_PCI_IOV +enum rtas_iov_fw_value_map { + NUM_RES_PROPERTY = 0, /* Number of Resources */ + LOW_INT = 1, /* Lowest 32 bits of Address */ + START_OF_ENTRIES = 2, /* Always start of entry */ + APERTURE_PROPERTY = 2, /* Start of entry+ to Aperture Size */ + WDW_SIZE_PROPERTY = 4, /* Start of entry+ to Window Size */ + NEXT_ENTRY = 7 /* Go to next entry on array */ +}; + +enum get_iov_fw_value_index { + BAR_ADDRS = 1, /* Get Bar Address */ + APERTURE_SIZE = 2, /* Get Aperture Size */ + WDW_SIZE = 3 /* Get Window Size */ +}; + +static resource_size_t pseries_get_iov_fw_value(struct pci_dev *dev, int resno, + enum get_iov_fw_value_index value) +{ + const int *indexes; + struct device_node *dn = pci_device_to_OF_node(dev); + int i, num_res, ret = 0; + + indexes = of_get_property(dn, "ibm,open-sriov-vf-bar-info", NULL); + if (!indexes) + return 0; + + /* + * First element in the array is the number of Bars + * returned. Search through the list to find the matching + * bar + */ + num_res = of_read_number(&indexes[NUM_RES_PROPERTY], 1); + if (resno >= num_res) + return 0; /* or an error */ + + i = START_OF_ENTRIES + NEXT_ENTRY * resno; + switch (value) { + case BAR_ADDRS: + ret = of_read_number(&indexes[i], 2); + break; + case APERTURE_SIZE: + ret = of_read_number(&indexes[i + APERTURE_PROPERTY], 2); + break; + case WDW_SIZE: + ret = of_read_number(&indexes[i + WDW_SIZE_PROPERTY], 2); + break; + } + + return ret; +} + +static void of_pci_set_vf_bar_size(struct pci_dev *dev, const int *indexes) +{ + struct resource *res; + resource_size_t base, size; + int i, r, num_res; + + num_res = of_read_number(&indexes[NUM_RES_PROPERTY], 1); + num_res = min_t(int, num_res, PCI_SRIOV_NUM_BARS); + for (i = START_OF_ENTRIES, r = 0; r < num_res && r < PCI_SRIOV_NUM_BARS; + i += NEXT_ENTRY, r++) { + res = &dev->resource[r + PCI_IOV_RESOURCES]; + base = of_read_number(&indexes[i], 2); + size = of_read_number(&indexes[i + APERTURE_PROPERTY], 2); + res->flags = pci_parse_of_flags(of_read_number + (&indexes[i + LOW_INT], 1), 0); + res->flags |= (IORESOURCE_MEM_64 | IORESOURCE_PCI_FIXED); + res->name = pci_name(dev); + res->start = base; + res->end = base + size - 1; + } +} + +static void of_pci_parse_iov_addrs(struct pci_dev *dev, const int *indexes) +{ + struct resource *res, *root, *conflict; + resource_size_t base, size; + int i, r, num_res; + + /* + * First element in the array is the number of Bars + * returned. Search through the list to find the matching + * bars assign them from firmware into resources structure. + */ + num_res = of_read_number(&indexes[NUM_RES_PROPERTY], 1); + for (i = START_OF_ENTRIES, r = 0; r < num_res && r < PCI_SRIOV_NUM_BARS; + i += NEXT_ENTRY, r++) { + res = &dev->resource[r + PCI_IOV_RESOURCES]; + base = of_read_number(&indexes[i], 2); + size = of_read_number(&indexes[i + WDW_SIZE_PROPERTY], 2); + res->name = pci_name(dev); + res->start = base; + res->end = base + size - 1; + root = &iomem_resource; + dev_dbg(&dev->dev, + "pSeries IOV BAR %d: trying firmware assignment %pR\n", + r + PCI_IOV_RESOURCES, res); + conflict = request_resource_conflict(root, res); + if (conflict) { + dev_info(&dev->dev, + "BAR %d: %pR conflicts with %s %pR\n", + r + PCI_IOV_RESOURCES, res, + conflict->name, conflict); + res->flags |= IORESOURCE_UNSET; + } + } +} + +static void pseries_disable_sriov_resources(struct pci_dev *pdev) +{ + int i; + + pci_warn(pdev, "No hypervisor support for SR-IOV on this device, IOV BARs disabled.\n"); + for (i = 0; i < PCI_SRIOV_NUM_BARS; i++) + pdev->resource[i + PCI_IOV_RESOURCES].flags = 0; +} + +static void pseries_pci_fixup_resources(struct pci_dev *pdev) +{ + const int *indexes; + struct device_node *dn = pci_device_to_OF_node(pdev); + + /*Firmware must support open sriov otherwise dont configure*/ + indexes = of_get_property(dn, "ibm,open-sriov-vf-bar-info", NULL); + if (indexes) + of_pci_set_vf_bar_size(pdev, indexes); + else + pseries_disable_sriov_resources(pdev); +} + +static void pseries_pci_fixup_iov_resources(struct pci_dev *pdev) +{ + const int *indexes; + struct device_node *dn = pci_device_to_OF_node(pdev); + + if (!pdev->is_physfn) + return; + /*Firmware must support open sriov otherwise don't configure*/ + indexes = of_get_property(dn, "ibm,open-sriov-vf-bar-info", NULL); + if (indexes) + of_pci_parse_iov_addrs(pdev, indexes); + else + pseries_disable_sriov_resources(pdev); +} + +static resource_size_t pseries_pci_iov_resource_alignment(struct pci_dev *pdev, + int resno) +{ + const __be32 *reg; + struct device_node *dn = pci_device_to_OF_node(pdev); + + /*Firmware must support open sriov otherwise report regular alignment*/ + reg = of_get_property(dn, "ibm,is-open-sriov-pf", NULL); + if (!reg) + return pci_iov_resource_size(pdev, resno); + + if (!pdev->is_physfn) + return 0; + return pseries_get_iov_fw_value(pdev, + resno - PCI_IOV_RESOURCES, + APERTURE_SIZE); +} +#endif + +static void __init pSeries_setup_arch(void) +{ + set_arch_panic_timeout(10, ARCH_PANIC_TIMEOUT); + + /* Discover PIC type and setup ppc_md accordingly */ + smp_init_pseries(); + + // Setup CPU hotplug callbacks + pseries_cpu_hotplug_init(); + + if (radix_enabled() && !mmu_has_feature(MMU_FTR_GTSE)) + if (!firmware_has_feature(FW_FEATURE_RPT_INVALIDATE)) + panic("BUG: Radix support requires either GTSE or RPT_INVALIDATE\n"); + + + /* openpic global configuration register (64-bit format). */ + /* openpic Interrupt Source Unit pointer (64-bit format). */ + /* python0 facility area (mmio) (64-bit format) REAL address. */ + + /* init to some ~sane value until calibrate_delay() runs */ + loops_per_jiffy = 50000000; + + fwnmi_init(); + + pseries_setup_security_mitigations(); + if (!radix_enabled()) + pseries_lpar_read_hblkrm_characteristics(); + + /* By default, only probe PCI (can be overridden by rtas_pci) */ + pci_add_flags(PCI_PROBE_ONLY); + + /* Find and initialize PCI host bridges */ + init_pci_config_tokens(); + of_reconfig_notifier_register(&pci_dn_reconfig_nb); + + pSeries_nvram_init(); + + if (firmware_has_feature(FW_FEATURE_LPAR)) { + vpa_init(boot_cpuid); + + if (lppaca_shared_proc()) { + static_branch_enable(&shared_processor); + pv_spinlocks_init(); +#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING + static_key_slow_inc(¶virt_steal_enabled); + if (steal_acc) + static_key_slow_inc(¶virt_steal_rq_enabled); +#endif + } + + ppc_md.power_save = pseries_lpar_idle; + ppc_md.enable_pmcs = pseries_lpar_enable_pmcs; +#ifdef CONFIG_PCI_IOV + ppc_md.pcibios_fixup_resources = + pseries_pci_fixup_resources; + ppc_md.pcibios_fixup_sriov = + pseries_pci_fixup_iov_resources; + ppc_md.pcibios_iov_resource_alignment = + pseries_pci_iov_resource_alignment; +#endif + } else { + /* No special idle routine */ + ppc_md.enable_pmcs = power4_enable_pmcs; + } + + ppc_md.pcibios_root_bridge_prepare = pseries_root_bridge_prepare; + pseries_rng_init(); +} + +static void pseries_panic(char *str) +{ + panic_flush_kmsg_end(); + rtas_os_term(str); +} + +static int __init pSeries_init_panel(void) +{ + /* Manually leave the kernel version on the panel. */ +#ifdef __BIG_ENDIAN__ + ppc_md.progress("Linux ppc64\n", 0); +#else + ppc_md.progress("Linux ppc64le\n", 0); +#endif + ppc_md.progress(init_utsname()->version, 0); + + return 0; +} +machine_arch_initcall(pseries, pSeries_init_panel); + +static int pseries_set_dabr(unsigned long dabr, unsigned long dabrx) +{ + return plpar_hcall_norets(H_SET_DABR, dabr); +} + +static int pseries_set_xdabr(unsigned long dabr, unsigned long dabrx) +{ + /* Have to set at least one bit in the DABRX according to PAPR */ + if (dabrx == 0 && dabr == 0) + dabrx = DABRX_USER; + /* PAPR says we can only set kernel and user bits */ + dabrx &= DABRX_KERNEL | DABRX_USER; + + return plpar_hcall_norets(H_SET_XDABR, dabr, dabrx); +} + +static int pseries_set_dawr(int nr, unsigned long dawr, unsigned long dawrx) +{ + /* PAPR says we can't set HYP */ + dawrx &= ~DAWRX_HYP; + + if (nr == 0) + return plpar_set_watchpoint0(dawr, dawrx); + else + return plpar_set_watchpoint1(dawr, dawrx); +} + +#define CMO_CHARACTERISTICS_TOKEN 44 +#define CMO_MAXLENGTH 1026 + +void pSeries_coalesce_init(void) +{ + struct hvcall_mpp_x_data mpp_x_data; + + if (firmware_has_feature(FW_FEATURE_CMO) && !h_get_mpp_x(&mpp_x_data)) + powerpc_firmware_features |= FW_FEATURE_XCMO; + else + powerpc_firmware_features &= ~FW_FEATURE_XCMO; +} + +/** + * fw_cmo_feature_init - FW_FEATURE_CMO is not stored in ibm,hypertas-functions, + * handle that here. (Stolen from parse_system_parameter_string) + */ +static void __init pSeries_cmo_feature_init(void) +{ + static struct papr_sysparm_buf buf __initdata; + static_assert(sizeof(buf.val) >= CMO_MAXLENGTH); + char *ptr, *key, *value, *end; + int page_order = IOMMU_PAGE_SHIFT_4K; + + pr_debug(" -> fw_cmo_feature_init()\n"); + + if (papr_sysparm_get(PAPR_SYSPARM_COOP_MEM_OVERCOMMIT_ATTRS, &buf)) { + pr_debug("CMO not available\n"); + pr_debug(" <- fw_cmo_feature_init()\n"); + return; + } + + end = &buf.val[CMO_MAXLENGTH]; + ptr = &buf.val[0]; + key = value = ptr; + + while (*ptr && (ptr <= end)) { + /* Separate the key and value by replacing '=' with '\0' and + * point the value at the string after the '=' + */ + if (ptr[0] == '=') { + ptr[0] = '\0'; + value = ptr + 1; + } else if (ptr[0] == '\0' || ptr[0] == ',') { + /* Terminate the string containing the key/value pair */ + ptr[0] = '\0'; + + if (key == value) { + pr_debug("Malformed key/value pair\n"); + /* Never found a '=', end processing */ + break; + } + + if (0 == strcmp(key, "CMOPageSize")) + page_order = simple_strtol(value, NULL, 10); + else if (0 == strcmp(key, "PrPSP")) + CMO_PrPSP = simple_strtol(value, NULL, 10); + else if (0 == strcmp(key, "SecPSP")) + CMO_SecPSP = simple_strtol(value, NULL, 10); + value = key = ptr + 1; + } + ptr++; + } + + /* Page size is returned as the power of 2 of the page size, + * convert to the page size in bytes before returning + */ + CMO_PageSize = 1 << page_order; + pr_debug("CMO_PageSize = %lu\n", CMO_PageSize); + + if (CMO_PrPSP != -1 || CMO_SecPSP != -1) { + pr_info("CMO enabled\n"); + pr_debug("CMO enabled, PrPSP=%d, SecPSP=%d\n", CMO_PrPSP, + CMO_SecPSP); + powerpc_firmware_features |= FW_FEATURE_CMO; + pSeries_coalesce_init(); + } else + pr_debug("CMO not enabled, PrPSP=%d, SecPSP=%d\n", CMO_PrPSP, + CMO_SecPSP); + pr_debug(" <- fw_cmo_feature_init()\n"); +} + +static void __init pseries_add_hw_description(void) +{ + struct device_node *dn; + const char *s; + + dn = of_find_node_by_path("/openprom"); + if (dn) { + if (of_property_read_string(dn, "model", &s) == 0) + seq_buf_printf(&ppc_hw_desc, "of:%s ", s); + + of_node_put(dn); + } + + dn = of_find_node_by_path("/hypervisor"); + if (dn) { + if (of_property_read_string(dn, "compatible", &s) == 0) + seq_buf_printf(&ppc_hw_desc, "hv:%s ", s); + + of_node_put(dn); + return; + } + + if (of_property_read_bool(of_root, "ibm,powervm-partition") || + of_property_read_bool(of_root, "ibm,fw-net-version")) + seq_buf_printf(&ppc_hw_desc, "hv:phyp "); +} + +/* + * Early initialization. Relocation is on but do not reference unbolted pages + */ +static void __init pseries_init(void) +{ + pr_debug(" -> pseries_init()\n"); + + pseries_add_hw_description(); + +#ifdef CONFIG_HVC_CONSOLE + if (firmware_has_feature(FW_FEATURE_LPAR)) + hvc_vio_init_early(); +#endif + if (firmware_has_feature(FW_FEATURE_XDABR)) + ppc_md.set_dabr = pseries_set_xdabr; + else if (firmware_has_feature(FW_FEATURE_DABR)) + ppc_md.set_dabr = pseries_set_dabr; + + if (firmware_has_feature(FW_FEATURE_SET_MODE)) + ppc_md.set_dawr = pseries_set_dawr; + + pSeries_cmo_feature_init(); + iommu_init_early_pSeries(); + + pr_debug(" <- pseries_init()\n"); +} + +/** + * pseries_power_off - tell firmware about how to power off the system. + * + * This function calls either the power-off rtas token in normal cases + * or the ibm,power-off-ups token (if present & requested) in case of + * a power failure. If power-off token is used, power on will only be + * possible with power button press. If ibm,power-off-ups token is used + * it will allow auto poweron after power is restored. + */ +static void pseries_power_off(void) +{ + int rc; + int rtas_poweroff_ups_token = rtas_function_token(RTAS_FN_IBM_POWER_OFF_UPS); + + if (rtas_flash_term_hook) + rtas_flash_term_hook(SYS_POWER_OFF); + + if (rtas_poweron_auto == 0 || + rtas_poweroff_ups_token == RTAS_UNKNOWN_SERVICE) { + rc = rtas_call(rtas_function_token(RTAS_FN_POWER_OFF), 2, 1, NULL, -1, -1); + printk(KERN_INFO "RTAS power-off returned %d\n", rc); + } else { + rc = rtas_call(rtas_poweroff_ups_token, 0, 1, NULL); + printk(KERN_INFO "RTAS ibm,power-off-ups returned %d\n", rc); + } + for (;;); +} + +static int __init pSeries_probe(void) +{ + if (!of_node_is_type(of_root, "chrp")) + return 0; + + /* Cell blades firmware claims to be chrp while it's not. Until this + * is fixed, we need to avoid those here. + */ + if (of_machine_is_compatible("IBM,CPBW-1.0") || + of_machine_is_compatible("IBM,CBEA")) + return 0; + + pm_power_off = pseries_power_off; + + pr_debug("Machine is%s LPAR !\n", + (powerpc_firmware_features & FW_FEATURE_LPAR) ? "" : " not"); + + pseries_init(); + + return 1; +} + +static int pSeries_pci_probe_mode(struct pci_bus *bus) +{ + if (firmware_has_feature(FW_FEATURE_LPAR)) + return PCI_PROBE_DEVTREE; + return PCI_PROBE_NORMAL; +} + +#ifdef CONFIG_MEMORY_HOTPLUG +static unsigned long pseries_memory_block_size(void) +{ + return memory_block_size; +} +#endif + +struct pci_controller_ops pseries_pci_controller_ops = { + .probe_mode = pSeries_pci_probe_mode, +#ifdef CONFIG_SPAPR_TCE_IOMMU + .device_group = pSeries_pci_device_group, +#endif +}; + +define_machine(pseries) { + .name = "pSeries", + .probe = pSeries_probe, + .setup_arch = pSeries_setup_arch, + .init_IRQ = pseries_init_irq, + .show_cpuinfo = pSeries_show_cpuinfo, + .log_error = pSeries_log_error, + .discover_phbs = pSeries_discover_phbs, + .pcibios_fixup = pSeries_final_fixup, + .restart = rtas_restart, + .halt = rtas_halt, + .panic = pseries_panic, + .get_boot_time = rtas_get_boot_time, + .get_rtc_time = rtas_get_rtc_time, + .set_rtc_time = rtas_set_rtc_time, + .progress = rtas_progress, + .system_reset_exception = pSeries_system_reset_exception, + .machine_check_early = pseries_machine_check_realmode, + .machine_check_exception = pSeries_machine_check_exception, + .machine_check_log_err = pSeries_machine_check_log_err, +#ifdef CONFIG_KEXEC_CORE + .machine_kexec = pseries_machine_kexec, + .kexec_cpu_down = pseries_kexec_cpu_down, +#endif +#ifdef CONFIG_MEMORY_HOTPLUG + .memory_block_size = pseries_memory_block_size, +#endif +}; diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c new file mode 100644 index 000000000..c597711ef --- /dev/null +++ b/arch/powerpc/platforms/pseries/smp.c @@ -0,0 +1,282 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * SMP support for pSeries machines. + * + * Dave Engebretsen, Peter Bergner, and + * Mike Corrigan {engebret|bergner|mikec}@us.ibm.com + * + * Plus various changes from other IBM teams... + */ + + +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/smp.h> +#include <linux/interrupt.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/spinlock.h> +#include <linux/cache.h> +#include <linux/err.h> +#include <linux/device.h> +#include <linux/cpu.h> +#include <linux/pgtable.h> + +#include <asm/ptrace.h> +#include <linux/atomic.h> +#include <asm/irq.h> +#include <asm/page.h> +#include <asm/io.h> +#include <asm/smp.h> +#include <asm/paca.h> +#include <asm/machdep.h> +#include <asm/cputable.h> +#include <asm/firmware.h> +#include <asm/rtas.h> +#include <asm/vdso_datapage.h> +#include <asm/cputhreads.h> +#include <asm/xics.h> +#include <asm/xive.h> +#include <asm/dbell.h> +#include <asm/plpar_wrappers.h> +#include <asm/code-patching.h> +#include <asm/svm.h> +#include <asm/kvm_guest.h> + +#include "pseries.h" + +/* + * The Primary thread of each non-boot processor was started from the OF client + * interface by prom_hold_cpus and is spinning on secondary_hold_spinloop. + */ +static cpumask_var_t of_spin_mask; + +/* Query where a cpu is now. Return codes #defined in plpar_wrappers.h */ +int smp_query_cpu_stopped(unsigned int pcpu) +{ + int cpu_status, status; + int qcss_tok = rtas_function_token(RTAS_FN_QUERY_CPU_STOPPED_STATE); + + if (qcss_tok == RTAS_UNKNOWN_SERVICE) { + printk_once(KERN_INFO + "Firmware doesn't support query-cpu-stopped-state\n"); + return QCSS_HARDWARE_ERROR; + } + + status = rtas_call(qcss_tok, 1, 2, &cpu_status, pcpu); + if (status != 0) { + printk(KERN_ERR + "RTAS query-cpu-stopped-state failed: %i\n", status); + return status; + } + + return cpu_status; +} + +/** + * smp_startup_cpu() - start the given cpu + * + * At boot time, there is nothing to do for primary threads which were + * started from Open Firmware. For anything else, call RTAS with the + * appropriate start location. + * + * Returns: + * 0 - failure + * 1 - success + */ +static inline int smp_startup_cpu(unsigned int lcpu) +{ + int status; + unsigned long start_here = + __pa(ppc_function_entry(generic_secondary_smp_init)); + unsigned int pcpu; + int start_cpu; + + if (cpumask_test_cpu(lcpu, of_spin_mask)) + /* Already started by OF and sitting in spin loop */ + return 1; + + pcpu = get_hard_smp_processor_id(lcpu); + + /* Check to see if the CPU out of FW already for kexec */ + if (smp_query_cpu_stopped(pcpu) == QCSS_NOT_STOPPED){ + cpumask_set_cpu(lcpu, of_spin_mask); + return 1; + } + + /* + * If the RTAS start-cpu token does not exist then presume the + * cpu is already spinning. + */ + start_cpu = rtas_function_token(RTAS_FN_START_CPU); + if (start_cpu == RTAS_UNKNOWN_SERVICE) + return 1; + + status = rtas_call(start_cpu, 3, 1, NULL, pcpu, start_here, pcpu); + if (status != 0) { + printk(KERN_ERR "start-cpu failed: %i\n", status); + return 0; + } + + return 1; +} + +static void smp_setup_cpu(int cpu) +{ + if (xive_enabled()) + xive_smp_setup_cpu(); + else if (cpu != boot_cpuid) + xics_setup_cpu(); + + if (firmware_has_feature(FW_FEATURE_SPLPAR)) + vpa_init(cpu); + + cpumask_clear_cpu(cpu, of_spin_mask); +} + +static int smp_pSeries_kick_cpu(int nr) +{ + if (nr < 0 || nr >= nr_cpu_ids) + return -EINVAL; + + if (!smp_startup_cpu(nr)) + return -ENOENT; + + /* + * The processor is currently spinning, waiting for the + * cpu_start field to become non-zero After we set cpu_start, + * the processor will continue on to secondary_start + */ + paca_ptrs[nr]->cpu_start = 1; + + return 0; +} + +static int pseries_smp_prepare_cpu(int cpu) +{ + if (xive_enabled()) + return xive_smp_prepare_cpu(cpu); + return 0; +} + +/* Cause IPI as setup by the interrupt controller (xics or xive) */ +static void (*ic_cause_ipi)(int cpu) __ro_after_init; + +/* Use msgsndp doorbells target is a sibling, else use interrupt controller */ +static void dbell_or_ic_cause_ipi(int cpu) +{ + if (doorbell_try_core_ipi(cpu)) + return; + + ic_cause_ipi(cpu); +} + +static int pseries_cause_nmi_ipi(int cpu) +{ + int hwcpu; + + if (cpu == NMI_IPI_ALL_OTHERS) { + hwcpu = H_SIGNAL_SYS_RESET_ALL_OTHERS; + } else { + if (cpu < 0) { + WARN_ONCE(true, "incorrect cpu parameter %d", cpu); + return 0; + } + + hwcpu = get_hard_smp_processor_id(cpu); + } + + if (plpar_signal_sys_reset(hwcpu) == H_SUCCESS) + return 1; + + return 0; +} + +static __init void pSeries_smp_probe(void) +{ + if (xive_enabled()) + xive_smp_probe(); + else + xics_smp_probe(); + + /* No doorbell facility, must use the interrupt controller for IPIs */ + if (!cpu_has_feature(CPU_FTR_DBELL)) + return; + + /* Doorbells can only be used for IPIs between SMT siblings */ + if (!cpu_has_feature(CPU_FTR_SMT)) + return; + + check_kvm_guest(); + + if (is_kvm_guest()) { + /* + * KVM emulates doorbells by disabling FSCR[MSGP] so msgsndp + * faults to the hypervisor which then reads the instruction + * from guest memory, which tends to be slower than using XIVE. + */ + if (xive_enabled()) + return; + + /* + * XICS hcalls aren't as fast, so we can use msgsndp (which + * also helps exercise KVM emulation), however KVM can't + * emulate secure guests because it can't read the instruction + * out of their memory. + */ + if (is_secure_guest()) + return; + } + + /* + * Under PowerVM, FSCR[MSGP] is enabled as guest vCPU siblings are + * gang scheduled on the same physical core, so doorbells are always + * faster than the interrupt controller, and they can be used by + * secure guests. + */ + + ic_cause_ipi = smp_ops->cause_ipi; + smp_ops->cause_ipi = dbell_or_ic_cause_ipi; +} + +static struct smp_ops_t pseries_smp_ops = { + .message_pass = NULL, /* Use smp_muxed_ipi_message_pass */ + .cause_ipi = NULL, /* Filled at runtime by pSeries_smp_probe() */ + .cause_nmi_ipi = pseries_cause_nmi_ipi, + .probe = pSeries_smp_probe, + .prepare_cpu = pseries_smp_prepare_cpu, + .kick_cpu = smp_pSeries_kick_cpu, + .setup_cpu = smp_setup_cpu, + .cpu_bootable = smp_generic_cpu_bootable, +}; + +/* This is called very early */ +void __init smp_init_pseries(void) +{ + int i; + + pr_debug(" -> smp_init_pSeries()\n"); + smp_ops = &pseries_smp_ops; + + alloc_bootmem_cpumask_var(&of_spin_mask); + + /* + * Mark threads which are still spinning in hold loops + * + * We know prom_init will not have started them if RTAS supports + * query-cpu-stopped-state. + */ + if (rtas_function_token(RTAS_FN_QUERY_CPU_STOPPED_STATE) == RTAS_UNKNOWN_SERVICE) { + if (cpu_has_feature(CPU_FTR_SMT)) { + for_each_present_cpu(i) { + if (cpu_thread_in_core(i) == 0) + cpumask_set_cpu(i, of_spin_mask); + } + } else + cpumask_copy(of_spin_mask, cpu_present_mask); + + cpumask_clear_cpu(boot_cpuid, of_spin_mask); + } + + pr_debug(" <- smp_init_pSeries()\n"); +} diff --git a/arch/powerpc/platforms/pseries/suspend.c b/arch/powerpc/platforms/pseries/suspend.c new file mode 100644 index 000000000..5c4343547 --- /dev/null +++ b/arch/powerpc/platforms/pseries/suspend.c @@ -0,0 +1,189 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2010 Brian King IBM Corporation + */ + +#include <linux/cpu.h> +#include <linux/delay.h> +#include <linux/suspend.h> +#include <linux/stat.h> +#include <asm/firmware.h> +#include <asm/hvcall.h> +#include <asm/machdep.h> +#include <asm/mmu.h> +#include <asm/rtas.h> +#include <asm/topology.h> + +static struct device suspend_dev; + +/** + * pseries_suspend_begin - First phase of hibernation + * + * Check to ensure we are in a valid state to hibernate + * + * Return value: + * 0 on success / other on failure + **/ +static int pseries_suspend_begin(u64 stream_id) +{ + long vasi_state, rc; + unsigned long retbuf[PLPAR_HCALL_BUFSIZE]; + + /* Make sure the state is valid */ + rc = plpar_hcall(H_VASI_STATE, retbuf, stream_id); + + vasi_state = retbuf[0]; + + if (rc) { + pr_err("pseries_suspend_begin: vasi_state returned %ld\n",rc); + return rc; + } else if (vasi_state == H_VASI_ENABLED) { + return -EAGAIN; + } else if (vasi_state != H_VASI_SUSPENDING) { + pr_err("pseries_suspend_begin: vasi_state returned state %ld\n", + vasi_state); + return -EIO; + } + return 0; +} + +/** + * pseries_suspend_enter - Final phase of hibernation + * + * Return value: + * 0 on success / other on failure + **/ +static int pseries_suspend_enter(suspend_state_t state) +{ + return rtas_ibm_suspend_me(NULL); +} + +/** + * store_hibernate - Initiate partition hibernation + * @dev: subsys root device + * @attr: device attribute struct + * @buf: buffer + * @count: buffer size + * + * Write the stream ID received from the HMC to this file + * to trigger hibernating the partition + * + * Return value: + * number of bytes printed to buffer / other on failure + **/ +static ssize_t store_hibernate(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + u64 stream_id; + int rc; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + stream_id = simple_strtoul(buf, NULL, 16); + + do { + rc = pseries_suspend_begin(stream_id); + if (rc == -EAGAIN) + ssleep(1); + } while (rc == -EAGAIN); + + if (!rc) + rc = pm_suspend(PM_SUSPEND_MEM); + + if (!rc) { + rc = count; + post_mobility_fixup(); + } + + + return rc; +} + +#define USER_DT_UPDATE 0 +#define KERN_DT_UPDATE 1 + +/** + * show_hibernate - Report device tree update responsibilty + * @dev: subsys root device + * @attr: device attribute struct + * @buf: buffer + * + * Report whether a device tree update is performed by the kernel after a + * resume, or if drmgr must coordinate the update from user space. + * + * Return value: + * 0 if drmgr is to initiate update, and 1 otherwise + **/ +static ssize_t show_hibernate(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + return sprintf(buf, "%d\n", KERN_DT_UPDATE); +} + +static DEVICE_ATTR(hibernate, 0644, show_hibernate, store_hibernate); + +static struct bus_type suspend_subsys = { + .name = "power", + .dev_name = "power", +}; + +static const struct platform_suspend_ops pseries_suspend_ops = { + .valid = suspend_valid_only_mem, + .enter = pseries_suspend_enter, +}; + +/** + * pseries_suspend_sysfs_register - Register with sysfs + * + * Return value: + * 0 on success / other on failure + **/ +static int pseries_suspend_sysfs_register(struct device *dev) +{ + struct device *dev_root; + int rc; + + if ((rc = subsys_system_register(&suspend_subsys, NULL))) + return rc; + + dev->id = 0; + dev->bus = &suspend_subsys; + + dev_root = bus_get_dev_root(&suspend_subsys); + if (dev_root) { + rc = device_create_file(dev_root, &dev_attr_hibernate); + put_device(dev_root); + if (rc) + goto subsys_unregister; + } + + return 0; + +subsys_unregister: + bus_unregister(&suspend_subsys); + return rc; +} + +/** + * pseries_suspend_init - initcall for pSeries suspend + * + * Return value: + * 0 on success / other on failure + **/ +static int __init pseries_suspend_init(void) +{ + int rc; + + if (!firmware_has_feature(FW_FEATURE_LPAR)) + return 0; + + if ((rc = pseries_suspend_sysfs_register(&suspend_dev))) + return rc; + + suspend_set_ops(&pseries_suspend_ops); + return 0; +} +machine_device_initcall(pseries, pseries_suspend_init); diff --git a/arch/powerpc/platforms/pseries/svm.c b/arch/powerpc/platforms/pseries/svm.c new file mode 100644 index 000000000..3b4045d50 --- /dev/null +++ b/arch/powerpc/platforms/pseries/svm.c @@ -0,0 +1,94 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Secure VM platform + * + * Copyright 2018 IBM Corporation + * Author: Anshuman Khandual <khandual@linux.vnet.ibm.com> + */ + +#include <linux/mm.h> +#include <linux/memblock.h> +#include <linux/cc_platform.h> +#include <asm/machdep.h> +#include <asm/svm.h> +#include <asm/swiotlb.h> +#include <asm/ultravisor.h> +#include <asm/dtl.h> + +static int __init init_svm(void) +{ + if (!is_secure_guest()) + return 0; + + /* Don't release the SWIOTLB buffer. */ + ppc_swiotlb_enable = 1; + + /* + * Since the guest memory is inaccessible to the host, devices always + * need to use the SWIOTLB buffer for DMA even if dma_capable() says + * otherwise. + */ + ppc_swiotlb_flags |= SWIOTLB_ANY | SWIOTLB_FORCE; + + /* Share the SWIOTLB buffer with the host. */ + swiotlb_update_mem_attributes(); + + return 0; +} +machine_early_initcall(pseries, init_svm); + +int set_memory_encrypted(unsigned long addr, int numpages) +{ + if (!cc_platform_has(CC_ATTR_MEM_ENCRYPT)) + return 0; + + if (!PAGE_ALIGNED(addr)) + return -EINVAL; + + uv_unshare_page(PHYS_PFN(__pa(addr)), numpages); + + return 0; +} + +int set_memory_decrypted(unsigned long addr, int numpages) +{ + if (!cc_platform_has(CC_ATTR_MEM_ENCRYPT)) + return 0; + + if (!PAGE_ALIGNED(addr)) + return -EINVAL; + + uv_share_page(PHYS_PFN(__pa(addr)), numpages); + + return 0; +} + +/* There's one dispatch log per CPU. */ +#define NR_DTL_PAGE (DISPATCH_LOG_BYTES * CONFIG_NR_CPUS / PAGE_SIZE) + +static struct page *dtl_page_store[NR_DTL_PAGE]; +static long dtl_nr_pages; + +static bool is_dtl_page_shared(struct page *page) +{ + long i; + + for (i = 0; i < dtl_nr_pages; i++) + if (dtl_page_store[i] == page) + return true; + + return false; +} + +void dtl_cache_ctor(void *addr) +{ + unsigned long pfn = PHYS_PFN(__pa(addr)); + struct page *page = pfn_to_page(pfn); + + if (!is_dtl_page_shared(page)) { + dtl_page_store[dtl_nr_pages] = page; + dtl_nr_pages++; + WARN_ON(dtl_nr_pages >= NR_DTL_PAGE); + uv_share_page(pfn, 1); + } +} diff --git a/arch/powerpc/platforms/pseries/vas-sysfs.c b/arch/powerpc/platforms/pseries/vas-sysfs.c new file mode 100644 index 000000000..f9f682724 --- /dev/null +++ b/arch/powerpc/platforms/pseries/vas-sysfs.c @@ -0,0 +1,281 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright 2022-23 IBM Corp. + */ + +#define pr_fmt(fmt) "vas: " fmt + +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/miscdevice.h> +#include <linux/kobject.h> +#include <linux/slab.h> +#include <linux/mm.h> + +#include "vas.h" + +#ifdef CONFIG_SYSFS +static struct kobject *pseries_vas_kobj; +static struct kobject *gzip_caps_kobj; + +struct vas_caps_entry { + struct kobject kobj; + struct vas_cop_feat_caps *caps; +}; + +#define to_caps_entry(entry) container_of(entry, struct vas_caps_entry, kobj) + +/* + * This function is used to get the notification from the drmgr when + * QoS credits are changed. + */ +static ssize_t update_total_credits_store(struct vas_cop_feat_caps *caps, + const char *buf, size_t count) +{ + int err; + u16 creds; + + err = kstrtou16(buf, 0, &creds); + /* + * The user space interface from the management console + * notifies OS with the new QoS credits and then the + * hypervisor. So OS has to use this new credits value + * and reconfigure VAS windows (close or reopen depends + * on the credits available) instead of depending on VAS + * QoS capabilities from the hypervisor. + */ + if (!err) + err = vas_reconfig_capabilties(caps->win_type, creds); + + if (err) + return -EINVAL; + + pr_info("Set QoS total credits %u\n", creds); + + return count; +} + +#define sysfs_caps_entry_read(_name) \ +static ssize_t _name##_show(struct vas_cop_feat_caps *caps, char *buf) \ +{ \ + return sprintf(buf, "%d\n", atomic_read(&caps->_name)); \ +} + +struct vas_sysfs_entry { + struct attribute attr; + ssize_t (*show)(struct vas_cop_feat_caps *, char *); + ssize_t (*store)(struct vas_cop_feat_caps *, const char *, size_t); +}; + +#define VAS_ATTR_RO(_name) \ + sysfs_caps_entry_read(_name); \ + static struct vas_sysfs_entry _name##_attribute = __ATTR(_name, \ + 0444, _name##_show, NULL); + +/* + * Create sysfs interface: + * /sys/devices/virtual/misc/vas/vas0/gzip/default_capabilities + * This directory contains the following VAS GZIP capabilities + * for the default credit type. + * /sys/devices/virtual/misc/vas/vas0/gzip/default_capabilities/nr_total_credits + * Total number of default credits assigned to the LPAR which + * can be changed with DLPAR operation. + * /sys/devices/virtual/misc/vas/vas0/gzip/default_capabilities/nr_used_credits + * Number of credits used by the user space. One credit will + * be assigned for each window open. + * + * /sys/devices/virtual/misc/vas/vas0/gzip/qos_capabilities + * This directory contains the following VAS GZIP capabilities + * for the Quality of Service (QoS) credit type. + * /sys/devices/virtual/misc/vas/vas0/gzip/qos_capabilities/nr_total_credits + * Total number of QoS credits assigned to the LPAR. The user + * has to define this value using HMC interface. It can be + * changed dynamically by the user. + * /sys/devices/virtual/misc/vas/vas0/gzip/qos_capabilities/nr_used_credits + * Number of credits used by the user space. + * /sys/devices/virtual/misc/vas/vas0/gzip/qos_capabilities/update_total_credits + * Update total QoS credits dynamically + */ + +VAS_ATTR_RO(nr_total_credits); +VAS_ATTR_RO(nr_used_credits); + +static struct vas_sysfs_entry update_total_credits_attribute = + __ATTR(update_total_credits, 0200, NULL, update_total_credits_store); + +static struct attribute *vas_def_capab_attrs[] = { + &nr_total_credits_attribute.attr, + &nr_used_credits_attribute.attr, + NULL, +}; +ATTRIBUTE_GROUPS(vas_def_capab); + +static struct attribute *vas_qos_capab_attrs[] = { + &nr_total_credits_attribute.attr, + &nr_used_credits_attribute.attr, + &update_total_credits_attribute.attr, + NULL, +}; +ATTRIBUTE_GROUPS(vas_qos_capab); + +static ssize_t vas_type_show(struct kobject *kobj, struct attribute *attr, + char *buf) +{ + struct vas_caps_entry *centry; + struct vas_cop_feat_caps *caps; + struct vas_sysfs_entry *entry; + + centry = to_caps_entry(kobj); + caps = centry->caps; + entry = container_of(attr, struct vas_sysfs_entry, attr); + + if (!entry->show) + return -EIO; + + return entry->show(caps, buf); +} + +static ssize_t vas_type_store(struct kobject *kobj, struct attribute *attr, + const char *buf, size_t count) +{ + struct vas_caps_entry *centry; + struct vas_cop_feat_caps *caps; + struct vas_sysfs_entry *entry; + + centry = to_caps_entry(kobj); + caps = centry->caps; + entry = container_of(attr, struct vas_sysfs_entry, attr); + if (!entry->store) + return -EIO; + + return entry->store(caps, buf, count); +} + +static void vas_type_release(struct kobject *kobj) +{ + struct vas_caps_entry *centry = to_caps_entry(kobj); + kfree(centry); +} + +static const struct sysfs_ops vas_sysfs_ops = { + .show = vas_type_show, + .store = vas_type_store, +}; + +static struct kobj_type vas_def_attr_type = { + .release = vas_type_release, + .sysfs_ops = &vas_sysfs_ops, + .default_groups = vas_def_capab_groups, +}; + +static struct kobj_type vas_qos_attr_type = { + .release = vas_type_release, + .sysfs_ops = &vas_sysfs_ops, + .default_groups = vas_qos_capab_groups, +}; + +static char *vas_caps_kobj_name(struct vas_caps_entry *centry, + struct kobject **kobj) +{ + struct vas_cop_feat_caps *caps = centry->caps; + + if (caps->descriptor == VAS_GZIP_QOS_CAPABILITIES) { + kobject_init(¢ry->kobj, &vas_qos_attr_type); + *kobj = gzip_caps_kobj; + return "qos_capabilities"; + } else if (caps->descriptor == VAS_GZIP_DEFAULT_CAPABILITIES) { + kobject_init(¢ry->kobj, &vas_def_attr_type); + *kobj = gzip_caps_kobj; + return "default_capabilities"; + } else + return "Unknown"; +} + +/* + * Add feature specific capability dir entry. + * Ex: VDefGzip or VQosGzip + */ +int sysfs_add_vas_caps(struct vas_cop_feat_caps *caps) +{ + struct vas_caps_entry *centry; + struct kobject *kobj = NULL; + int ret = 0; + char *name; + + centry = kzalloc(sizeof(*centry), GFP_KERNEL); + if (!centry) + return -ENOMEM; + + centry->caps = caps; + name = vas_caps_kobj_name(centry, &kobj); + + if (kobj) { + ret = kobject_add(¢ry->kobj, kobj, "%s", name); + + if (ret) { + pr_err("VAS: sysfs kobject add / event failed %d\n", + ret); + kobject_put(¢ry->kobj); + } + } + + return ret; +} + +static struct miscdevice vas_miscdev = { + .minor = MISC_DYNAMIC_MINOR, + .name = "vas", +}; + +/* + * Add VAS and VasCaps (overall capabilities) dir entries. + */ +int __init sysfs_pseries_vas_init(struct vas_all_caps *vas_caps) +{ + int ret; + + ret = misc_register(&vas_miscdev); + if (ret < 0) { + pr_err("%s: register vas misc device failed\n", __func__); + return ret; + } + + /* + * The hypervisor does not expose multiple VAS instances, but can + * see multiple VAS instances on PowerNV. So create 'vas0' directory + * on pseries. + */ + pseries_vas_kobj = kobject_create_and_add("vas0", + &vas_miscdev.this_device->kobj); + if (!pseries_vas_kobj) { + misc_deregister(&vas_miscdev); + pr_err("Failed to create VAS sysfs entry\n"); + return -ENOMEM; + } + + if ((vas_caps->feat_type & VAS_GZIP_QOS_FEAT_BIT) || + (vas_caps->feat_type & VAS_GZIP_DEF_FEAT_BIT)) { + gzip_caps_kobj = kobject_create_and_add("gzip", + pseries_vas_kobj); + if (!gzip_caps_kobj) { + pr_err("Failed to create VAS GZIP capability entry\n"); + kobject_put(pseries_vas_kobj); + misc_deregister(&vas_miscdev); + return -ENOMEM; + } + } + + return 0; +} + +#else +int sysfs_add_vas_caps(struct vas_cop_feat_caps *caps) +{ + return 0; +} + +int __init sysfs_pseries_vas_init(struct vas_all_caps *vas_caps) +{ + return 0; +} +#endif diff --git a/arch/powerpc/platforms/pseries/vas.c b/arch/powerpc/platforms/pseries/vas.c new file mode 100644 index 000000000..71d52a670 --- /dev/null +++ b/arch/powerpc/platforms/pseries/vas.c @@ -0,0 +1,1121 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright 2020-21 IBM Corp. + */ + +#define pr_fmt(fmt) "vas: " fmt + +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/export.h> +#include <linux/types.h> +#include <linux/delay.h> +#include <linux/slab.h> +#include <linux/interrupt.h> +#include <linux/irqdomain.h> +#include <asm/machdep.h> +#include <asm/hvcall.h> +#include <asm/plpar_wrappers.h> +#include <asm/firmware.h> +#include <asm/vphn.h> +#include <asm/vas.h> +#include "vas.h" + +#define VAS_INVALID_WIN_ADDRESS 0xFFFFFFFFFFFFFFFFul +#define VAS_DEFAULT_DOMAIN_ID 0xFFFFFFFFFFFFFFFFul +/* The hypervisor allows one credit per window right now */ +#define DEF_WIN_CREDS 1 + +static struct vas_all_caps caps_all; +static bool copypaste_feat; +static struct hv_vas_cop_feat_caps hv_cop_caps; + +static struct vas_caps vascaps[VAS_MAX_FEAT_TYPE]; +static DEFINE_MUTEX(vas_pseries_mutex); +static bool migration_in_progress; + +static long hcall_return_busy_check(long rc) +{ + /* Check if we are stalled for some time */ + if (H_IS_LONG_BUSY(rc)) { + msleep(get_longbusy_msecs(rc)); + rc = H_BUSY; + } else if (rc == H_BUSY) { + cond_resched(); + } + + return rc; +} + +/* + * Allocate VAS window hcall + */ +static int h_allocate_vas_window(struct pseries_vas_window *win, u64 *domain, + u8 wintype, u16 credits) +{ + long retbuf[PLPAR_HCALL9_BUFSIZE] = {0}; + long rc; + + do { + rc = plpar_hcall9(H_ALLOCATE_VAS_WINDOW, retbuf, wintype, + credits, domain[0], domain[1], domain[2], + domain[3], domain[4], domain[5]); + + rc = hcall_return_busy_check(rc); + } while (rc == H_BUSY); + + if (rc == H_SUCCESS) { + if (win->win_addr == VAS_INVALID_WIN_ADDRESS) { + pr_err("H_ALLOCATE_VAS_WINDOW: COPY/PASTE is not supported\n"); + return -ENOTSUPP; + } + win->vas_win.winid = retbuf[0]; + win->win_addr = retbuf[1]; + win->complete_irq = retbuf[2]; + win->fault_irq = retbuf[3]; + return 0; + } + + pr_err("H_ALLOCATE_VAS_WINDOW error: %ld, wintype: %u, credits: %u\n", + rc, wintype, credits); + + return -EIO; +} + +/* + * Deallocate VAS window hcall. + */ +static int h_deallocate_vas_window(u64 winid) +{ + long rc; + + do { + rc = plpar_hcall_norets(H_DEALLOCATE_VAS_WINDOW, winid); + + rc = hcall_return_busy_check(rc); + } while (rc == H_BUSY); + + if (rc == H_SUCCESS) + return 0; + + pr_err("H_DEALLOCATE_VAS_WINDOW error: %ld, winid: %llu\n", + rc, winid); + return -EIO; +} + +/* + * Modify VAS window. + * After the window is opened with allocate window hcall, configure it + * with flags and LPAR PID before using. + */ +static int h_modify_vas_window(struct pseries_vas_window *win) +{ + long rc; + + /* + * AMR value is not supported in Linux VAS implementation. + * The hypervisor ignores it if 0 is passed. + */ + do { + rc = plpar_hcall_norets(H_MODIFY_VAS_WINDOW, + win->vas_win.winid, win->pid, 0, + VAS_MOD_WIN_FLAGS, 0); + + rc = hcall_return_busy_check(rc); + } while (rc == H_BUSY); + + if (rc == H_SUCCESS) + return 0; + + pr_err("H_MODIFY_VAS_WINDOW error: %ld, winid %u pid %u\n", + rc, win->vas_win.winid, win->pid); + return -EIO; +} + +/* + * This hcall is used to determine the capabilities from the hypervisor. + * @hcall: H_QUERY_VAS_CAPABILITIES or H_QUERY_NX_CAPABILITIES + * @query_type: If 0 is passed, the hypervisor returns the overall + * capabilities which provides all feature(s) that are + * available. Then query the hypervisor to get the + * corresponding capabilities for the specific feature. + * Example: H_QUERY_VAS_CAPABILITIES provides VAS GZIP QoS + * and VAS GZIP Default capabilities. + * H_QUERY_NX_CAPABILITIES provides NX GZIP + * capabilities. + * @result: Return buffer to save capabilities. + */ +int h_query_vas_capabilities(const u64 hcall, u8 query_type, u64 result) +{ + long rc; + + rc = plpar_hcall_norets(hcall, query_type, result); + + if (rc == H_SUCCESS) + return 0; + + /* H_FUNCTION means HV does not support VAS so don't print an error */ + if (rc != H_FUNCTION) { + pr_err("%s error %ld, query_type %u, result buffer 0x%llx\n", + (hcall == H_QUERY_VAS_CAPABILITIES) ? + "H_QUERY_VAS_CAPABILITIES" : + "H_QUERY_NX_CAPABILITIES", + rc, query_type, result); + } + + return -EIO; +} +EXPORT_SYMBOL_GPL(h_query_vas_capabilities); + +/* + * hcall to get fault CRB from the hypervisor. + */ +static int h_get_nx_fault(u32 winid, u64 buffer) +{ + long rc; + + rc = plpar_hcall_norets(H_GET_NX_FAULT, winid, buffer); + + if (rc == H_SUCCESS) + return 0; + + pr_err("H_GET_NX_FAULT error: %ld, winid %u, buffer 0x%llx\n", + rc, winid, buffer); + return -EIO; + +} + +/* + * Handle the fault interrupt. + * When the fault interrupt is received for each window, query the + * hypervisor to get the fault CRB on the specific fault. Then + * process the CRB by updating CSB or send signal if the user space + * CSB is invalid. + * Note: The hypervisor forwards an interrupt for each fault request. + * So one fault CRB to process for each H_GET_NX_FAULT hcall. + */ +static irqreturn_t pseries_vas_fault_thread_fn(int irq, void *data) +{ + struct pseries_vas_window *txwin = data; + struct coprocessor_request_block crb; + struct vas_user_win_ref *tsk_ref; + int rc; + + while (atomic_read(&txwin->pending_faults)) { + rc = h_get_nx_fault(txwin->vas_win.winid, (u64)virt_to_phys(&crb)); + if (!rc) { + tsk_ref = &txwin->vas_win.task_ref; + vas_dump_crb(&crb); + vas_update_csb(&crb, tsk_ref); + } + atomic_dec(&txwin->pending_faults); + } + + return IRQ_HANDLED; +} + +/* + * irq_default_primary_handler() can be used only with IRQF_ONESHOT + * which disables IRQ before executing the thread handler and enables + * it after. But this disabling interrupt sets the VAS IRQ OFF + * state in the hypervisor. If the NX generates fault interrupt + * during this window, the hypervisor will not deliver this + * interrupt to the LPAR. So use VAS specific IRQ handler instead + * of calling the default primary handler. + */ +static irqreturn_t pseries_vas_irq_handler(int irq, void *data) +{ + struct pseries_vas_window *txwin = data; + + /* + * The thread hanlder will process this interrupt if it is + * already running. + */ + atomic_inc(&txwin->pending_faults); + + return IRQ_WAKE_THREAD; +} + +/* + * Allocate window and setup IRQ mapping. + */ +static int allocate_setup_window(struct pseries_vas_window *txwin, + u64 *domain, u8 wintype) +{ + int rc; + + rc = h_allocate_vas_window(txwin, domain, wintype, DEF_WIN_CREDS); + if (rc) + return rc; + /* + * On PowerVM, the hypervisor setup and forwards the fault + * interrupt per window. So the IRQ setup and fault handling + * will be done for each open window separately. + */ + txwin->fault_virq = irq_create_mapping(NULL, txwin->fault_irq); + if (!txwin->fault_virq) { + pr_err("Failed irq mapping %d\n", txwin->fault_irq); + rc = -EINVAL; + goto out_win; + } + + txwin->name = kasprintf(GFP_KERNEL, "vas-win-%d", + txwin->vas_win.winid); + if (!txwin->name) { + rc = -ENOMEM; + goto out_irq; + } + + rc = request_threaded_irq(txwin->fault_virq, + pseries_vas_irq_handler, + pseries_vas_fault_thread_fn, 0, + txwin->name, txwin); + if (rc) { + pr_err("VAS-Window[%d]: Request IRQ(%u) failed with %d\n", + txwin->vas_win.winid, txwin->fault_virq, rc); + goto out_free; + } + + txwin->vas_win.wcreds_max = DEF_WIN_CREDS; + + return 0; +out_free: + kfree(txwin->name); +out_irq: + irq_dispose_mapping(txwin->fault_virq); +out_win: + h_deallocate_vas_window(txwin->vas_win.winid); + return rc; +} + +static inline void free_irq_setup(struct pseries_vas_window *txwin) +{ + free_irq(txwin->fault_virq, txwin); + kfree(txwin->name); + irq_dispose_mapping(txwin->fault_virq); +} + +static struct vas_window *vas_allocate_window(int vas_id, u64 flags, + enum vas_cop_type cop_type) +{ + long domain[PLPAR_HCALL9_BUFSIZE] = {VAS_DEFAULT_DOMAIN_ID}; + struct vas_cop_feat_caps *cop_feat_caps; + struct vas_caps *caps; + struct pseries_vas_window *txwin; + int rc; + + txwin = kzalloc(sizeof(*txwin), GFP_KERNEL); + if (!txwin) + return ERR_PTR(-ENOMEM); + + /* + * A VAS window can have many credits which means that many + * requests can be issued simultaneously. But the hypervisor + * restricts one credit per window. + * The hypervisor introduces 2 different types of credits: + * Default credit type (Uses normal priority FIFO): + * A limited number of credits are assigned to partitions + * based on processor entitlement. But these credits may be + * over-committed on a system depends on whether the CPUs + * are in shared or dedicated modes - that is, more requests + * may be issued across the system than NX can service at + * once which can result in paste command failure (RMA_busy). + * Then the process has to resend requests or fall-back to + * SW compression. + * Quality of Service (QoS) credit type (Uses high priority FIFO): + * To avoid NX HW contention, the system admins can assign + * QoS credits for each LPAR so that this partition is + * guaranteed access to NX resources. These credits are + * assigned to partitions via the HMC. + * Refer PAPR for more information. + * + * Allocate window with QoS credits if user requested. Otherwise + * default credits are used. + */ + if (flags & VAS_TX_WIN_FLAG_QOS_CREDIT) + caps = &vascaps[VAS_GZIP_QOS_FEAT_TYPE]; + else + caps = &vascaps[VAS_GZIP_DEF_FEAT_TYPE]; + + cop_feat_caps = &caps->caps; + + if (atomic_inc_return(&cop_feat_caps->nr_used_credits) > + atomic_read(&cop_feat_caps->nr_total_credits)) { + pr_err_ratelimited("Credits are not available to allocate window\n"); + rc = -EINVAL; + goto out; + } + + if (vas_id == -1) { + /* + * The user space is requesting to allocate a window on + * a VAS instance where the process is executing. + * On PowerVM, domain values are passed to the hypervisor + * to select VAS instance. Useful if the process is + * affinity to NUMA node. + * The hypervisor selects VAS instance if + * VAS_DEFAULT_DOMAIN_ID (-1) is passed for domain values. + * The h_allocate_vas_window hcall is defined to take a + * domain values as specified by h_home_node_associativity, + * So no unpacking needs to be done. + */ + rc = plpar_hcall9(H_HOME_NODE_ASSOCIATIVITY, domain, + VPHN_FLAG_VCPU, hard_smp_processor_id()); + if (rc != H_SUCCESS) { + pr_err("H_HOME_NODE_ASSOCIATIVITY error: %d\n", rc); + goto out; + } + } + + txwin->pid = mfspr(SPRN_PID); + + /* + * Allocate / Deallocate window hcalls and setup / free IRQs + * have to be protected with mutex. + * Open VAS window: Allocate window hcall and setup IRQ + * Close VAS window: Deallocate window hcall and free IRQ + * The hypervisor waits until all NX requests are + * completed before closing the window. So expects OS + * to handle NX faults, means IRQ can be freed only + * after the deallocate window hcall is returned. + * So once the window is closed with deallocate hcall before + * the IRQ is freed, it can be assigned to new allocate + * hcall with the same fault IRQ by the hypervisor. It can + * result in setup IRQ fail for the new window since the + * same fault IRQ is not freed by the OS before. + */ + mutex_lock(&vas_pseries_mutex); + if (migration_in_progress) { + rc = -EBUSY; + } else { + rc = allocate_setup_window(txwin, (u64 *)&domain[0], + cop_feat_caps->win_type); + if (!rc) + caps->nr_open_wins_progress++; + } + + mutex_unlock(&vas_pseries_mutex); + if (rc) + goto out; + + /* + * Modify window and it is ready to use. + */ + rc = h_modify_vas_window(txwin); + if (!rc) + rc = get_vas_user_win_ref(&txwin->vas_win.task_ref); + if (rc) + goto out_free; + + txwin->win_type = cop_feat_caps->win_type; + + /* + * The migration SUSPEND thread sets migration_in_progress and + * closes all open windows from the list. But the window is + * added to the list after open and modify HCALLs. So possible + * that migration_in_progress is set before modify HCALL which + * may cause some windows are still open when the hypervisor + * initiates the migration. + * So checks the migration_in_progress flag again and close all + * open windows. + * + * Possible to lose the acquired credit with DLPAR core + * removal after the window is opened. So if there are any + * closed windows (means with lost credits), do not give new + * window to user space. New windows will be opened only + * after the existing windows are reopened when credits are + * available. + */ + mutex_lock(&vas_pseries_mutex); + if (!caps->nr_close_wins && !migration_in_progress) { + list_add(&txwin->win_list, &caps->list); + caps->nr_open_windows++; + caps->nr_open_wins_progress--; + mutex_unlock(&vas_pseries_mutex); + vas_user_win_add_mm_context(&txwin->vas_win.task_ref); + return &txwin->vas_win; + } + mutex_unlock(&vas_pseries_mutex); + + put_vas_user_win_ref(&txwin->vas_win.task_ref); + rc = -EBUSY; + pr_err_ratelimited("No credit is available to allocate window\n"); + +out_free: + /* + * Window is not operational. Free IRQ before closing + * window so that do not have to hold mutex. + */ + free_irq_setup(txwin); + h_deallocate_vas_window(txwin->vas_win.winid); + /* + * Hold mutex and reduce nr_open_wins_progress counter. + */ + mutex_lock(&vas_pseries_mutex); + caps->nr_open_wins_progress--; + mutex_unlock(&vas_pseries_mutex); +out: + atomic_dec(&cop_feat_caps->nr_used_credits); + kfree(txwin); + return ERR_PTR(rc); +} + +static u64 vas_paste_address(struct vas_window *vwin) +{ + struct pseries_vas_window *win; + + win = container_of(vwin, struct pseries_vas_window, vas_win); + return win->win_addr; +} + +static int deallocate_free_window(struct pseries_vas_window *win) +{ + int rc = 0; + + /* + * The hypervisor waits for all requests including faults + * are processed before closing the window - Means all + * credits have to be returned. In the case of fault + * request, a credit is returned after OS issues + * H_GET_NX_FAULT hcall. + * So free IRQ after executing H_DEALLOCATE_VAS_WINDOW + * hcall. + */ + rc = h_deallocate_vas_window(win->vas_win.winid); + if (!rc) + free_irq_setup(win); + + return rc; +} + +static int vas_deallocate_window(struct vas_window *vwin) +{ + struct pseries_vas_window *win; + struct vas_cop_feat_caps *caps; + int rc = 0; + + if (!vwin) + return -EINVAL; + + win = container_of(vwin, struct pseries_vas_window, vas_win); + + /* Should not happen */ + if (win->win_type >= VAS_MAX_FEAT_TYPE) { + pr_err("Window (%u): Invalid window type %u\n", + vwin->winid, win->win_type); + return -EINVAL; + } + + caps = &vascaps[win->win_type].caps; + mutex_lock(&vas_pseries_mutex); + /* + * VAS window is already closed in the hypervisor when + * lost the credit or with migration. So just remove the entry + * from the list, remove task references and free vas_window + * struct. + */ + if (!(win->vas_win.status & VAS_WIN_NO_CRED_CLOSE) && + !(win->vas_win.status & VAS_WIN_MIGRATE_CLOSE)) { + rc = deallocate_free_window(win); + if (rc) { + mutex_unlock(&vas_pseries_mutex); + return rc; + } + } else + vascaps[win->win_type].nr_close_wins--; + + list_del(&win->win_list); + atomic_dec(&caps->nr_used_credits); + vascaps[win->win_type].nr_open_windows--; + mutex_unlock(&vas_pseries_mutex); + + mm_context_remove_vas_window(vwin->task_ref.mm); + put_vas_user_win_ref(&vwin->task_ref); + + kfree(win); + return 0; +} + +static const struct vas_user_win_ops vops_pseries = { + .open_win = vas_allocate_window, /* Open and configure window */ + .paste_addr = vas_paste_address, /* To do copy/paste */ + .close_win = vas_deallocate_window, /* Close window */ +}; + +/* + * Supporting only nx-gzip coprocessor type now, but this API code + * extended to other coprocessor types later. + */ +int vas_register_api_pseries(struct module *mod, enum vas_cop_type cop_type, + const char *name) +{ + if (!copypaste_feat) + return -ENOTSUPP; + + return vas_register_coproc_api(mod, cop_type, name, &vops_pseries); +} +EXPORT_SYMBOL_GPL(vas_register_api_pseries); + +void vas_unregister_api_pseries(void) +{ + vas_unregister_coproc_api(); +} +EXPORT_SYMBOL_GPL(vas_unregister_api_pseries); + +/* + * Get the specific capabilities based on the feature type. + * Right now supports GZIP default and GZIP QoS capabilities. + */ +static int __init get_vas_capabilities(u8 feat, enum vas_cop_feat_type type, + struct hv_vas_cop_feat_caps *hv_caps) +{ + struct vas_cop_feat_caps *caps; + struct vas_caps *vcaps; + int rc = 0; + + vcaps = &vascaps[type]; + memset(vcaps, 0, sizeof(*vcaps)); + INIT_LIST_HEAD(&vcaps->list); + + vcaps->feat = feat; + caps = &vcaps->caps; + + rc = h_query_vas_capabilities(H_QUERY_VAS_CAPABILITIES, feat, + (u64)virt_to_phys(hv_caps)); + if (rc) + return rc; + + caps->user_mode = hv_caps->user_mode; + if (!(caps->user_mode & VAS_COPY_PASTE_USER_MODE)) { + pr_err("User space COPY/PASTE is not supported\n"); + return -ENOTSUPP; + } + + caps->descriptor = be64_to_cpu(hv_caps->descriptor); + caps->win_type = hv_caps->win_type; + if (caps->win_type >= VAS_MAX_FEAT_TYPE) { + pr_err("Unsupported window type %u\n", caps->win_type); + return -EINVAL; + } + caps->max_lpar_creds = be16_to_cpu(hv_caps->max_lpar_creds); + caps->max_win_creds = be16_to_cpu(hv_caps->max_win_creds); + atomic_set(&caps->nr_total_credits, + be16_to_cpu(hv_caps->target_lpar_creds)); + if (feat == VAS_GZIP_DEF_FEAT) { + caps->def_lpar_creds = be16_to_cpu(hv_caps->def_lpar_creds); + + if (caps->max_win_creds < DEF_WIN_CREDS) { + pr_err("Window creds(%u) > max allowed window creds(%u)\n", + DEF_WIN_CREDS, caps->max_win_creds); + return -EINVAL; + } + } + + rc = sysfs_add_vas_caps(caps); + if (rc) + return rc; + + copypaste_feat = true; + + return 0; +} + +/* + * VAS windows can be closed due to lost credits when the core is + * removed. So reopen them if credits are available due to DLPAR + * core add and set the window active status. When NX sees the page + * fault on the unmapped paste address, the kernel handles the fault + * by setting the remapping to new paste address if the window is + * active. + */ +static int reconfig_open_windows(struct vas_caps *vcaps, int creds, + bool migrate) +{ + long domain[PLPAR_HCALL9_BUFSIZE] = {VAS_DEFAULT_DOMAIN_ID}; + struct vas_cop_feat_caps *caps = &vcaps->caps; + struct pseries_vas_window *win = NULL, *tmp; + int rc, mv_ents = 0; + int flag; + + /* + * Nothing to do if there are no closed windows. + */ + if (!vcaps->nr_close_wins) + return 0; + + /* + * For the core removal, the hypervisor reduces the credits + * assigned to the LPAR and the kernel closes VAS windows + * in the hypervisor depends on reduced credits. The kernel + * uses LIFO (the last windows that are opened will be closed + * first) and expects to open in the same order when credits + * are available. + * For example, 40 windows are closed when the LPAR lost 2 cores + * (dedicated). If 1 core is added, this LPAR can have 20 more + * credits. It means the kernel can reopen 20 windows. So move + * 20 entries in the VAS windows lost and reopen next 20 windows. + * For partition migration, reopen all windows that are closed + * during resume. + */ + if ((vcaps->nr_close_wins > creds) && !migrate) + mv_ents = vcaps->nr_close_wins - creds; + + list_for_each_entry_safe(win, tmp, &vcaps->list, win_list) { + if (!mv_ents) + break; + + mv_ents--; + } + + /* + * Open windows if they are closed only with migration or + * DLPAR (lost credit) before. + */ + if (migrate) + flag = VAS_WIN_MIGRATE_CLOSE; + else + flag = VAS_WIN_NO_CRED_CLOSE; + + list_for_each_entry_safe_from(win, tmp, &vcaps->list, win_list) { + /* + * This window is closed with DLPAR and migration events. + * So reopen the window with the last event. + * The user space is not suspended with the current + * migration notifier. So the user space can issue DLPAR + * CPU hotplug while migration in progress. In this case + * this window will be opened with the last event. + */ + if ((win->vas_win.status & VAS_WIN_NO_CRED_CLOSE) && + (win->vas_win.status & VAS_WIN_MIGRATE_CLOSE)) { + win->vas_win.status &= ~flag; + continue; + } + + /* + * Nothing to do on this window if it is not closed + * with this flag + */ + if (!(win->vas_win.status & flag)) + continue; + + rc = allocate_setup_window(win, (u64 *)&domain[0], + caps->win_type); + if (rc) + return rc; + + rc = h_modify_vas_window(win); + if (rc) + goto out; + + mutex_lock(&win->vas_win.task_ref.mmap_mutex); + /* + * Set window status to active + */ + win->vas_win.status &= ~flag; + mutex_unlock(&win->vas_win.task_ref.mmap_mutex); + win->win_type = caps->win_type; + if (!--vcaps->nr_close_wins) + break; + } + + return 0; +out: + /* + * Window modify HCALL failed. So close the window to the + * hypervisor and return. + */ + free_irq_setup(win); + h_deallocate_vas_window(win->vas_win.winid); + return rc; +} + +/* + * The hypervisor reduces the available credits if the LPAR lost core. It + * means the excessive windows should not be active and the user space + * should not be using these windows to send compression requests to NX. + * So the kernel closes the excessive windows and unmap the paste address + * such that the user space receives paste instruction failure. Then up to + * the user space to fall back to SW compression and manage with the + * existing windows. + */ +static int reconfig_close_windows(struct vas_caps *vcap, int excess_creds, + bool migrate) +{ + struct pseries_vas_window *win, *tmp; + struct vas_user_win_ref *task_ref; + struct vm_area_struct *vma; + int rc = 0, flag; + + if (migrate) + flag = VAS_WIN_MIGRATE_CLOSE; + else + flag = VAS_WIN_NO_CRED_CLOSE; + + list_for_each_entry_safe(win, tmp, &vcap->list, win_list) { + /* + * This window is already closed due to lost credit + * or for migration before. Go for next window. + * For migration, nothing to do since this window + * closed for DLPAR and will be reopened even on + * the destination system with other DLPAR operation. + */ + if ((win->vas_win.status & VAS_WIN_MIGRATE_CLOSE) || + (win->vas_win.status & VAS_WIN_NO_CRED_CLOSE)) { + win->vas_win.status |= flag; + continue; + } + + task_ref = &win->vas_win.task_ref; + /* + * VAS mmap (coproc_mmap()) and its fault handler + * (vas_mmap_fault()) are called after holding mmap lock. + * So hold mmap mutex after mmap_lock to avoid deadlock. + */ + mmap_write_lock(task_ref->mm); + mutex_lock(&task_ref->mmap_mutex); + vma = task_ref->vma; + /* + * Number of available credits are reduced, So select + * and close windows. + */ + win->vas_win.status |= flag; + + /* + * vma is set in the original mapping. But this mapping + * is done with mmap() after the window is opened with ioctl. + * so we may not see the original mapping if the core remove + * is done before the original mmap() and after the ioctl. + */ + if (vma) + zap_vma_pages(vma); + + mutex_unlock(&task_ref->mmap_mutex); + mmap_write_unlock(task_ref->mm); + /* + * Close VAS window in the hypervisor, but do not + * free vas_window struct since it may be reused + * when the credit is available later (DLPAR with + * adding cores). This struct will be used + * later when the process issued with close(FD). + */ + rc = deallocate_free_window(win); + /* + * This failure is from the hypervisor. + * No way to stop migration for these failures. + * So ignore error and continue closing other windows. + */ + if (rc && !migrate) + return rc; + + vcap->nr_close_wins++; + + /* + * For migration, do not depend on lpar_creds in case if + * mismatch with the hypervisor value (should not happen). + * So close all active windows in the list and will be + * reopened windows based on the new lpar_creds on the + * destination system during resume. + */ + if (!migrate && !--excess_creds) + break; + } + + return 0; +} + +/* + * Get new VAS capabilities when the core add/removal configuration + * changes. Reconfig window configurations based on the credits + * availability from this new capabilities. + */ +int vas_reconfig_capabilties(u8 type, int new_nr_creds) +{ + struct vas_cop_feat_caps *caps; + int old_nr_creds; + struct vas_caps *vcaps; + int rc = 0, nr_active_wins; + + if (type >= VAS_MAX_FEAT_TYPE) { + pr_err("Invalid credit type %d\n", type); + return -EINVAL; + } + + vcaps = &vascaps[type]; + caps = &vcaps->caps; + + mutex_lock(&vas_pseries_mutex); + + old_nr_creds = atomic_read(&caps->nr_total_credits); + + atomic_set(&caps->nr_total_credits, new_nr_creds); + /* + * The total number of available credits may be decreased or + * increased with DLPAR operation. Means some windows have to be + * closed / reopened. Hold the vas_pseries_mutex so that the + * user space can not open new windows. + */ + if (old_nr_creds < new_nr_creds) { + /* + * If the existing target credits is less than the new + * target, reopen windows if they are closed due to + * the previous DLPAR (core removal). + */ + rc = reconfig_open_windows(vcaps, new_nr_creds - old_nr_creds, + false); + } else { + /* + * # active windows is more than new LPAR available + * credits. So close the excessive windows. + * On pseries, each window will have 1 credit. + */ + nr_active_wins = vcaps->nr_open_windows - vcaps->nr_close_wins; + if (nr_active_wins > new_nr_creds) + rc = reconfig_close_windows(vcaps, + nr_active_wins - new_nr_creds, + false); + } + + mutex_unlock(&vas_pseries_mutex); + return rc; +} + +int pseries_vas_dlpar_cpu(void) +{ + int new_nr_creds, rc; + + /* + * NX-GZIP is not enabled. Nothing to do for DLPAR event + */ + if (!copypaste_feat) + return 0; + + + rc = h_query_vas_capabilities(H_QUERY_VAS_CAPABILITIES, + vascaps[VAS_GZIP_DEF_FEAT_TYPE].feat, + (u64)virt_to_phys(&hv_cop_caps)); + if (!rc) { + new_nr_creds = be16_to_cpu(hv_cop_caps.target_lpar_creds); + rc = vas_reconfig_capabilties(VAS_GZIP_DEF_FEAT_TYPE, new_nr_creds); + } + + if (rc) + pr_err("Failed reconfig VAS capabilities with DLPAR\n"); + + return rc; +} + +/* + * Total number of default credits available (target_credits) + * in LPAR depends on number of cores configured. It varies based on + * whether processors are in shared mode or dedicated mode. + * Get the notifier when CPU configuration is changed with DLPAR + * operation so that get the new target_credits (vas default capabilities) + * and then update the existing windows usage if needed. + */ +static int pseries_vas_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct of_reconfig_data *rd = data; + struct device_node *dn = rd->dn; + const __be32 *intserv = NULL; + int len; + + /* + * For shared CPU partition, the hypervisor assigns total credits + * based on entitled core capacity. So updating VAS windows will + * be called from lparcfg_write(). + */ + if (is_shared_processor()) + return NOTIFY_OK; + + if ((action == OF_RECONFIG_ATTACH_NODE) || + (action == OF_RECONFIG_DETACH_NODE)) + intserv = of_get_property(dn, "ibm,ppc-interrupt-server#s", + &len); + /* + * Processor config is not changed + */ + if (!intserv) + return NOTIFY_OK; + + return pseries_vas_dlpar_cpu(); +} + +static struct notifier_block pseries_vas_nb = { + .notifier_call = pseries_vas_notifier, +}; + +/* + * For LPM, all windows have to be closed on the source partition + * before migration and reopen them on the destination partition + * after migration. So closing windows during suspend and + * reopen them during resume. + */ +int vas_migration_handler(int action) +{ + struct vas_cop_feat_caps *caps; + int old_nr_creds, new_nr_creds = 0; + struct vas_caps *vcaps; + int i, rc = 0; + + pr_info("VAS migration event %d\n", action); + + /* + * NX-GZIP is not enabled. Nothing to do for migration. + */ + if (!copypaste_feat) + return rc; + + if (action == VAS_SUSPEND) + migration_in_progress = true; + else + migration_in_progress = false; + + for (i = 0; i < VAS_MAX_FEAT_TYPE; i++) { + vcaps = &vascaps[i]; + caps = &vcaps->caps; + old_nr_creds = atomic_read(&caps->nr_total_credits); + + rc = h_query_vas_capabilities(H_QUERY_VAS_CAPABILITIES, + vcaps->feat, + (u64)virt_to_phys(&hv_cop_caps)); + if (!rc) { + new_nr_creds = be16_to_cpu(hv_cop_caps.target_lpar_creds); + /* + * Should not happen. But incase print messages, close + * all windows in the list during suspend and reopen + * windows based on new lpar_creds on the destination + * system. + */ + if (old_nr_creds != new_nr_creds) { + pr_err("Target credits mismatch with the hypervisor\n"); + pr_err("state(%d): lpar creds: %d HV lpar creds: %d\n", + action, old_nr_creds, new_nr_creds); + pr_err("Used creds: %d, Active creds: %d\n", + atomic_read(&caps->nr_used_credits), + vcaps->nr_open_windows - vcaps->nr_close_wins); + } + } else { + pr_err("state(%d): Get VAS capabilities failed with %d\n", + action, rc); + /* + * We can not stop migration with the current lpm + * implementation. So continue closing all windows in + * the list (during suspend) and return without + * opening windows (during resume) if VAS capabilities + * HCALL failed. + */ + if (action == VAS_RESUME) + goto out; + } + + switch (action) { + case VAS_SUSPEND: + mutex_lock(&vas_pseries_mutex); + rc = reconfig_close_windows(vcaps, vcaps->nr_open_windows, + true); + /* + * Windows are included in the list after successful + * open. So wait for closing these in-progress open + * windows in vas_allocate_window() which will be + * done if the migration_in_progress is set. + */ + while (vcaps->nr_open_wins_progress) { + mutex_unlock(&vas_pseries_mutex); + msleep(10); + mutex_lock(&vas_pseries_mutex); + } + mutex_unlock(&vas_pseries_mutex); + break; + case VAS_RESUME: + mutex_lock(&vas_pseries_mutex); + atomic_set(&caps->nr_total_credits, new_nr_creds); + rc = reconfig_open_windows(vcaps, new_nr_creds, true); + mutex_unlock(&vas_pseries_mutex); + break; + default: + /* should not happen */ + pr_err("Invalid migration action %d\n", action); + rc = -EINVAL; + goto out; + } + + /* + * Ignore errors during suspend and return for resume. + */ + if (rc && (action == VAS_RESUME)) + goto out; + } + + pr_info("VAS migration event (%d) successful\n", action); + +out: + return rc; +} + +static int __init pseries_vas_init(void) +{ + struct hv_vas_all_caps *hv_caps; + int rc = 0; + + /* + * Linux supports user space COPY/PASTE only with Radix + */ + if (!radix_enabled()) { + copypaste_feat = false; + pr_err("API is supported only with radix page tables\n"); + return -ENOTSUPP; + } + + hv_caps = kmalloc(sizeof(*hv_caps), GFP_KERNEL); + if (!hv_caps) + return -ENOMEM; + /* + * Get VAS overall capabilities by passing 0 to feature type. + */ + rc = h_query_vas_capabilities(H_QUERY_VAS_CAPABILITIES, 0, + (u64)virt_to_phys(hv_caps)); + if (rc) + goto out; + + caps_all.descriptor = be64_to_cpu(hv_caps->descriptor); + caps_all.feat_type = be64_to_cpu(hv_caps->feat_type); + + sysfs_pseries_vas_init(&caps_all); + + /* + * QOS capabilities available + */ + if (caps_all.feat_type & VAS_GZIP_QOS_FEAT_BIT) { + rc = get_vas_capabilities(VAS_GZIP_QOS_FEAT, + VAS_GZIP_QOS_FEAT_TYPE, &hv_cop_caps); + + if (rc) + goto out; + } + /* + * Default capabilities available + */ + if (caps_all.feat_type & VAS_GZIP_DEF_FEAT_BIT) + rc = get_vas_capabilities(VAS_GZIP_DEF_FEAT, + VAS_GZIP_DEF_FEAT_TYPE, &hv_cop_caps); + + if (!rc && copypaste_feat) { + if (firmware_has_feature(FW_FEATURE_LPAR)) + of_reconfig_notifier_register(&pseries_vas_nb); + + pr_info("GZIP feature is available\n"); + } else { + /* + * Should not happen, but only when get default + * capabilities HCALL failed. So disable copy paste + * feature. + */ + copypaste_feat = false; + } + +out: + kfree(hv_caps); + return rc; +} +machine_device_initcall(pseries, pseries_vas_init); diff --git a/arch/powerpc/platforms/pseries/vas.h b/arch/powerpc/platforms/pseries/vas.h new file mode 100644 index 000000000..45567cd13 --- /dev/null +++ b/arch/powerpc/platforms/pseries/vas.h @@ -0,0 +1,157 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Copyright 2020-21 IBM Corp. + */ + +#ifndef _VAS_H +#define _VAS_H +#include <asm/vas.h> +#include <linux/mutex.h> +#include <linux/stringify.h> + +/* + * VAS window modify flags + */ +#define VAS_MOD_WIN_CLOSE PPC_BIT(0) +#define VAS_MOD_WIN_JOBS_KILL PPC_BIT(1) +#define VAS_MOD_WIN_DR PPC_BIT(3) +#define VAS_MOD_WIN_PR PPC_BIT(4) +#define VAS_MOD_WIN_SF PPC_BIT(5) +#define VAS_MOD_WIN_TA PPC_BIT(6) +#define VAS_MOD_WIN_FLAGS (VAS_MOD_WIN_JOBS_KILL | VAS_MOD_WIN_DR | \ + VAS_MOD_WIN_PR | VAS_MOD_WIN_SF) + +#define VAS_WIN_ACTIVE 0x0 +#define VAS_WIN_CLOSED 0x1 +#define VAS_WIN_INACTIVE 0x2 /* Inactive due to HW failure */ +/* Process of being modified, deallocated, or quiesced */ +#define VAS_WIN_MOD_IN_PROCESS 0x3 + +#define VAS_COPY_PASTE_USER_MODE 0x00000001 +#define VAS_COP_OP_USER_MODE 0x00000010 + +#define VAS_GZIP_QOS_CAPABILITIES 0x56516F73477A6970 +#define VAS_GZIP_DEFAULT_CAPABILITIES 0x56446566477A6970 + +enum vas_migrate_action { + VAS_SUSPEND, + VAS_RESUME, +}; + +/* + * Co-processor feature - GZIP QoS windows or GZIP default windows + */ +enum vas_cop_feat_type { + VAS_GZIP_QOS_FEAT_TYPE, + VAS_GZIP_DEF_FEAT_TYPE, + VAS_MAX_FEAT_TYPE, +}; + +/* + * Use to get feature specific capabilities from the + * hypervisor. + */ +struct hv_vas_cop_feat_caps { + __be64 descriptor; + u8 win_type; /* Default or QoS type */ + u8 user_mode; + __be16 max_lpar_creds; + __be16 max_win_creds; + union { + __be16 reserved; + __be16 def_lpar_creds; /* Used for default capabilities */ + }; + __be16 target_lpar_creds; +} __packed __aligned(0x1000); + +/* + * Feature specific (QoS or default) capabilities. + */ +struct vas_cop_feat_caps { + u64 descriptor; + u8 win_type; /* Default or QoS type */ + u8 user_mode; /* User mode copy/paste or COP HCALL */ + u16 max_lpar_creds; /* Max credits available in LPAR */ + /* Max credits can be assigned per window */ + u16 max_win_creds; + union { + u16 reserved; /* Used for QoS credit type */ + u16 def_lpar_creds; /* Used for default credit type */ + }; + /* Total LPAR available credits. Can be different from max LPAR */ + /* credits due to DLPAR operation */ + atomic_t nr_total_credits; /* Total credits assigned to LPAR */ + atomic_t nr_used_credits; /* Used credits so far */ +}; + +/* + * Feature (QoS or Default) specific to store capabilities and + * the list of open windows. + */ +struct vas_caps { + struct vas_cop_feat_caps caps; + struct list_head list; /* List of open windows */ + int nr_open_wins_progress; /* Number of open windows in */ + /* progress. Used in migration */ + int nr_close_wins; /* closed windows in the hypervisor for DLPAR */ + int nr_open_windows; /* Number of successful open windows */ + u8 feat; /* Feature type */ +}; + +/* + * To get window information from the hypervisor. + */ +struct hv_vas_win_lpar { + __be16 version; + u8 win_type; + u8 status; + __be16 credits; /* No of credits assigned to this window */ + __be16 reserved; + __be32 pid; /* LPAR Process ID */ + __be32 tid; /* LPAR Thread ID */ + __be64 win_addr; /* Paste address */ + __be32 interrupt; /* Interrupt when NX request completes */ + __be32 fault; /* Interrupt when NX sees fault */ + /* Associativity Domain Identifiers as returned in */ + /* H_HOME_NODE_ASSOCIATIVITY */ + __be64 domain[6]; + __be64 win_util; /* Number of bytes processed */ +} __packed __aligned(0x1000); + +struct pseries_vas_window { + struct vas_window vas_win; + u64 win_addr; /* Physical paste address */ + u8 win_type; /* QoS or Default window */ + u32 complete_irq; /* Completion interrupt */ + u32 fault_irq; /* Fault interrupt */ + u64 domain[6]; /* Associativity domain Ids */ + /* this window is allocated */ + u64 util; + u32 pid; /* PID associated with this window */ + + /* List of windows opened which is used for LPM */ + struct list_head win_list; + u64 flags; + char *name; + int fault_virq; + atomic_t pending_faults; /* Number of pending faults */ +}; + +int sysfs_add_vas_caps(struct vas_cop_feat_caps *caps); +int vas_reconfig_capabilties(u8 type, int new_nr_creds); +int __init sysfs_pseries_vas_init(struct vas_all_caps *vas_caps); + +#ifdef CONFIG_PPC_VAS +int vas_migration_handler(int action); +int pseries_vas_dlpar_cpu(void); +#else +static inline int vas_migration_handler(int action) +{ + return 0; +} +static inline int pseries_vas_dlpar_cpu(void) +{ + return 0; +} +#endif +#endif /* _VAS_H */ diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c new file mode 100644 index 000000000..2dc9cbc4b --- /dev/null +++ b/arch/powerpc/platforms/pseries/vio.c @@ -0,0 +1,1729 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * IBM PowerPC Virtual I/O Infrastructure Support. + * + * Copyright (c) 2003,2008 IBM Corp. + * Dave Engebretsen engebret@us.ibm.com + * Santiago Leon santil@us.ibm.com + * Hollis Blanchard <hollisb@us.ibm.com> + * Stephen Rothwell + * Robert Jennings <rcjenn@us.ibm.com> + */ + +#include <linux/cpu.h> +#include <linux/types.h> +#include <linux/delay.h> +#include <linux/stat.h> +#include <linux/device.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/console.h> +#include <linux/export.h> +#include <linux/mm.h> +#include <linux/dma-map-ops.h> +#include <linux/kobject.h> +#include <linux/kexec.h> +#include <linux/of_irq.h> + +#include <asm/iommu.h> +#include <asm/dma.h> +#include <asm/vio.h> +#include <asm/prom.h> +#include <asm/firmware.h> +#include <asm/tce.h> +#include <asm/page.h> +#include <asm/hvcall.h> +#include <asm/machdep.h> + +static struct vio_dev vio_bus_device = { /* fake "parent" device */ + .name = "vio", + .type = "", + .dev.init_name = "vio", + .dev.bus = &vio_bus_type, +}; + +#ifdef CONFIG_PPC_SMLPAR +/** + * vio_cmo_pool - A pool of IO memory for CMO use + * + * @size: The size of the pool in bytes + * @free: The amount of free memory in the pool + */ +struct vio_cmo_pool { + size_t size; + size_t free; +}; + +/* How many ms to delay queued balance work */ +#define VIO_CMO_BALANCE_DELAY 100 + +/* Portion out IO memory to CMO devices by this chunk size */ +#define VIO_CMO_BALANCE_CHUNK 131072 + +/** + * vio_cmo_dev_entry - A device that is CMO-enabled and requires entitlement + * + * @vio_dev: struct vio_dev pointer + * @list: pointer to other devices on bus that are being tracked + */ +struct vio_cmo_dev_entry { + struct vio_dev *viodev; + struct list_head list; +}; + +/** + * vio_cmo - VIO bus accounting structure for CMO entitlement + * + * @lock: spinlock for entire structure + * @balance_q: work queue for balancing system entitlement + * @device_list: list of CMO-enabled devices requiring entitlement + * @entitled: total system entitlement in bytes + * @reserve: pool of memory from which devices reserve entitlement, incl. spare + * @excess: pool of excess entitlement not needed for device reserves or spare + * @spare: IO memory for device hotplug functionality + * @min: minimum necessary for system operation + * @desired: desired memory for system operation + * @curr: bytes currently allocated + * @high: high water mark for IO data usage + */ +static struct vio_cmo { + spinlock_t lock; + struct delayed_work balance_q; + struct list_head device_list; + size_t entitled; + struct vio_cmo_pool reserve; + struct vio_cmo_pool excess; + size_t spare; + size_t min; + size_t desired; + size_t curr; + size_t high; +} vio_cmo; + +/** + * vio_cmo_OF_devices - Count the number of OF devices that have DMA windows + */ +static int vio_cmo_num_OF_devs(void) +{ + struct device_node *node_vroot; + int count = 0; + + /* + * Count the number of vdevice entries with an + * ibm,my-dma-window OF property + */ + node_vroot = of_find_node_by_name(NULL, "vdevice"); + if (node_vroot) { + struct device_node *of_node; + struct property *prop; + + for_each_child_of_node(node_vroot, of_node) { + prop = of_find_property(of_node, "ibm,my-dma-window", + NULL); + if (prop) + count++; + } + } + of_node_put(node_vroot); + return count; +} + +/** + * vio_cmo_alloc - allocate IO memory for CMO-enable devices + * + * @viodev: VIO device requesting IO memory + * @size: size of allocation requested + * + * Allocations come from memory reserved for the devices and any excess + * IO memory available to all devices. The spare pool used to service + * hotplug must be equal to %VIO_CMO_MIN_ENT for the excess pool to be + * made available. + * + * Return codes: + * 0 for successful allocation and -ENOMEM for a failure + */ +static inline int vio_cmo_alloc(struct vio_dev *viodev, size_t size) +{ + unsigned long flags; + size_t reserve_free = 0; + size_t excess_free = 0; + int ret = -ENOMEM; + + spin_lock_irqsave(&vio_cmo.lock, flags); + + /* Determine the amount of free entitlement available in reserve */ + if (viodev->cmo.entitled > viodev->cmo.allocated) + reserve_free = viodev->cmo.entitled - viodev->cmo.allocated; + + /* If spare is not fulfilled, the excess pool can not be used. */ + if (vio_cmo.spare >= VIO_CMO_MIN_ENT) + excess_free = vio_cmo.excess.free; + + /* The request can be satisfied */ + if ((reserve_free + excess_free) >= size) { + vio_cmo.curr += size; + if (vio_cmo.curr > vio_cmo.high) + vio_cmo.high = vio_cmo.curr; + viodev->cmo.allocated += size; + size -= min(reserve_free, size); + vio_cmo.excess.free -= size; + ret = 0; + } + + spin_unlock_irqrestore(&vio_cmo.lock, flags); + return ret; +} + +/** + * vio_cmo_dealloc - deallocate IO memory from CMO-enable devices + * @viodev: VIO device freeing IO memory + * @size: size of deallocation + * + * IO memory is freed by the device back to the correct memory pools. + * The spare pool is replenished first from either memory pool, then + * the reserve pool is used to reduce device entitlement, the excess + * pool is used to increase the reserve pool toward the desired entitlement + * target, and then the remaining memory is returned to the pools. + * + */ +static inline void vio_cmo_dealloc(struct vio_dev *viodev, size_t size) +{ + unsigned long flags; + size_t spare_needed = 0; + size_t excess_freed = 0; + size_t reserve_freed = size; + size_t tmp; + int balance = 0; + + spin_lock_irqsave(&vio_cmo.lock, flags); + vio_cmo.curr -= size; + + /* Amount of memory freed from the excess pool */ + if (viodev->cmo.allocated > viodev->cmo.entitled) { + excess_freed = min(reserve_freed, (viodev->cmo.allocated - + viodev->cmo.entitled)); + reserve_freed -= excess_freed; + } + + /* Remove allocation from device */ + viodev->cmo.allocated -= (reserve_freed + excess_freed); + + /* Spare is a subset of the reserve pool, replenish it first. */ + spare_needed = VIO_CMO_MIN_ENT - vio_cmo.spare; + + /* + * Replenish the spare in the reserve pool from the excess pool. + * This moves entitlement into the reserve pool. + */ + if (spare_needed && excess_freed) { + tmp = min(excess_freed, spare_needed); + vio_cmo.excess.size -= tmp; + vio_cmo.reserve.size += tmp; + vio_cmo.spare += tmp; + excess_freed -= tmp; + spare_needed -= tmp; + balance = 1; + } + + /* + * Replenish the spare in the reserve pool from the reserve pool. + * This removes entitlement from the device down to VIO_CMO_MIN_ENT, + * if needed, and gives it to the spare pool. The amount of used + * memory in this pool does not change. + */ + if (spare_needed && reserve_freed) { + tmp = min3(spare_needed, reserve_freed, (viodev->cmo.entitled - VIO_CMO_MIN_ENT)); + + vio_cmo.spare += tmp; + viodev->cmo.entitled -= tmp; + reserve_freed -= tmp; + spare_needed -= tmp; + balance = 1; + } + + /* + * Increase the reserve pool until the desired allocation is met. + * Move an allocation freed from the excess pool into the reserve + * pool and schedule a balance operation. + */ + if (excess_freed && (vio_cmo.desired > vio_cmo.reserve.size)) { + tmp = min(excess_freed, (vio_cmo.desired - vio_cmo.reserve.size)); + + vio_cmo.excess.size -= tmp; + vio_cmo.reserve.size += tmp; + excess_freed -= tmp; + balance = 1; + } + + /* Return memory from the excess pool to that pool */ + if (excess_freed) + vio_cmo.excess.free += excess_freed; + + if (balance) + schedule_delayed_work(&vio_cmo.balance_q, VIO_CMO_BALANCE_DELAY); + spin_unlock_irqrestore(&vio_cmo.lock, flags); +} + +/** + * vio_cmo_entitlement_update - Manage system entitlement changes + * + * @new_entitlement: new system entitlement to attempt to accommodate + * + * Increases in entitlement will be used to fulfill the spare entitlement + * and the rest is given to the excess pool. Decreases, if they are + * possible, come from the excess pool and from unused device entitlement + * + * Returns: 0 on success, -ENOMEM when change can not be made + */ +int vio_cmo_entitlement_update(size_t new_entitlement) +{ + struct vio_dev *viodev; + struct vio_cmo_dev_entry *dev_ent; + unsigned long flags; + size_t avail, delta, tmp; + + spin_lock_irqsave(&vio_cmo.lock, flags); + + /* Entitlement increases */ + if (new_entitlement > vio_cmo.entitled) { + delta = new_entitlement - vio_cmo.entitled; + + /* Fulfill spare allocation */ + if (vio_cmo.spare < VIO_CMO_MIN_ENT) { + tmp = min(delta, (VIO_CMO_MIN_ENT - vio_cmo.spare)); + vio_cmo.spare += tmp; + vio_cmo.reserve.size += tmp; + delta -= tmp; + } + + /* Remaining new allocation goes to the excess pool */ + vio_cmo.entitled += delta; + vio_cmo.excess.size += delta; + vio_cmo.excess.free += delta; + + goto out; + } + + /* Entitlement decreases */ + delta = vio_cmo.entitled - new_entitlement; + avail = vio_cmo.excess.free; + + /* + * Need to check how much unused entitlement each device can + * sacrifice to fulfill entitlement change. + */ + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) { + if (avail >= delta) + break; + + viodev = dev_ent->viodev; + if ((viodev->cmo.entitled > viodev->cmo.allocated) && + (viodev->cmo.entitled > VIO_CMO_MIN_ENT)) + avail += viodev->cmo.entitled - + max_t(size_t, viodev->cmo.allocated, + VIO_CMO_MIN_ENT); + } + + if (delta <= avail) { + vio_cmo.entitled -= delta; + + /* Take entitlement from the excess pool first */ + tmp = min(vio_cmo.excess.free, delta); + vio_cmo.excess.size -= tmp; + vio_cmo.excess.free -= tmp; + delta -= tmp; + + /* + * Remove all but VIO_CMO_MIN_ENT bytes from devices + * until entitlement change is served + */ + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) { + if (!delta) + break; + + viodev = dev_ent->viodev; + tmp = 0; + if ((viodev->cmo.entitled > viodev->cmo.allocated) && + (viodev->cmo.entitled > VIO_CMO_MIN_ENT)) + tmp = viodev->cmo.entitled - + max_t(size_t, viodev->cmo.allocated, + VIO_CMO_MIN_ENT); + viodev->cmo.entitled -= min(tmp, delta); + delta -= min(tmp, delta); + } + } else { + spin_unlock_irqrestore(&vio_cmo.lock, flags); + return -ENOMEM; + } + +out: + schedule_delayed_work(&vio_cmo.balance_q, 0); + spin_unlock_irqrestore(&vio_cmo.lock, flags); + return 0; +} + +/** + * vio_cmo_balance - Balance entitlement among devices + * + * @work: work queue structure for this operation + * + * Any system entitlement above the minimum needed for devices, or + * already allocated to devices, can be distributed to the devices. + * The list of devices is iterated through to recalculate the desired + * entitlement level and to determine how much entitlement above the + * minimum entitlement is allocated to devices. + * + * Small chunks of the available entitlement are given to devices until + * their requirements are fulfilled or there is no entitlement left to give. + * Upon completion sizes of the reserve and excess pools are calculated. + * + * The system minimum entitlement level is also recalculated here. + * Entitlement will be reserved for devices even after vio_bus_remove to + * accommodate reloading the driver. The OF tree is walked to count the + * number of devices present and this will remove entitlement for devices + * that have actually left the system after having vio_bus_remove called. + */ +static void vio_cmo_balance(struct work_struct *work) +{ + struct vio_cmo *cmo; + struct vio_dev *viodev; + struct vio_cmo_dev_entry *dev_ent; + unsigned long flags; + size_t avail = 0, level, chunk, need; + int devcount = 0, fulfilled; + + cmo = container_of(work, struct vio_cmo, balance_q.work); + + spin_lock_irqsave(&vio_cmo.lock, flags); + + /* Calculate minimum entitlement and fulfill spare */ + cmo->min = vio_cmo_num_OF_devs() * VIO_CMO_MIN_ENT; + BUG_ON(cmo->min > cmo->entitled); + cmo->spare = min_t(size_t, VIO_CMO_MIN_ENT, (cmo->entitled - cmo->min)); + cmo->min += cmo->spare; + cmo->desired = cmo->min; + + /* + * Determine how much entitlement is available and reset device + * entitlements + */ + avail = cmo->entitled - cmo->spare; + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) { + viodev = dev_ent->viodev; + devcount++; + viodev->cmo.entitled = VIO_CMO_MIN_ENT; + cmo->desired += (viodev->cmo.desired - VIO_CMO_MIN_ENT); + avail -= max_t(size_t, viodev->cmo.allocated, VIO_CMO_MIN_ENT); + } + + /* + * Having provided each device with the minimum entitlement, loop + * over the devices portioning out the remaining entitlement + * until there is nothing left. + */ + level = VIO_CMO_MIN_ENT; + while (avail) { + fulfilled = 0; + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) { + viodev = dev_ent->viodev; + + if (viodev->cmo.desired <= level) { + fulfilled++; + continue; + } + + /* + * Give the device up to VIO_CMO_BALANCE_CHUNK + * bytes of entitlement, but do not exceed the + * desired level of entitlement for the device. + */ + chunk = min_t(size_t, avail, VIO_CMO_BALANCE_CHUNK); + chunk = min(chunk, (viodev->cmo.desired - + viodev->cmo.entitled)); + viodev->cmo.entitled += chunk; + + /* + * If the memory for this entitlement increase was + * already allocated to the device it does not come + * from the available pool being portioned out. + */ + need = max(viodev->cmo.allocated, viodev->cmo.entitled)- + max(viodev->cmo.allocated, level); + avail -= need; + + } + if (fulfilled == devcount) + break; + level += VIO_CMO_BALANCE_CHUNK; + } + + /* Calculate new reserve and excess pool sizes */ + cmo->reserve.size = cmo->min; + cmo->excess.free = 0; + cmo->excess.size = 0; + need = 0; + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) { + viodev = dev_ent->viodev; + /* Calculated reserve size above the minimum entitlement */ + if (viodev->cmo.entitled) + cmo->reserve.size += (viodev->cmo.entitled - + VIO_CMO_MIN_ENT); + /* Calculated used excess entitlement */ + if (viodev->cmo.allocated > viodev->cmo.entitled) + need += viodev->cmo.allocated - viodev->cmo.entitled; + } + cmo->excess.size = cmo->entitled - cmo->reserve.size; + cmo->excess.free = cmo->excess.size - need; + + cancel_delayed_work(to_delayed_work(work)); + spin_unlock_irqrestore(&vio_cmo.lock, flags); +} + +static void *vio_dma_iommu_alloc_coherent(struct device *dev, size_t size, + dma_addr_t *dma_handle, gfp_t flag, + unsigned long attrs) +{ + struct vio_dev *viodev = to_vio_dev(dev); + void *ret; + + if (vio_cmo_alloc(viodev, roundup(size, PAGE_SIZE))) { + atomic_inc(&viodev->cmo.allocs_failed); + return NULL; + } + + ret = iommu_alloc_coherent(dev, get_iommu_table_base(dev), size, + dma_handle, dev->coherent_dma_mask, flag, + dev_to_node(dev)); + if (unlikely(ret == NULL)) { + vio_cmo_dealloc(viodev, roundup(size, PAGE_SIZE)); + atomic_inc(&viodev->cmo.allocs_failed); + } + + return ret; +} + +static void vio_dma_iommu_free_coherent(struct device *dev, size_t size, + void *vaddr, dma_addr_t dma_handle, + unsigned long attrs) +{ + struct vio_dev *viodev = to_vio_dev(dev); + + iommu_free_coherent(get_iommu_table_base(dev), size, vaddr, dma_handle); + vio_cmo_dealloc(viodev, roundup(size, PAGE_SIZE)); +} + +static dma_addr_t vio_dma_iommu_map_page(struct device *dev, struct page *page, + unsigned long offset, size_t size, + enum dma_data_direction direction, + unsigned long attrs) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct iommu_table *tbl = get_iommu_table_base(dev); + dma_addr_t ret = DMA_MAPPING_ERROR; + + if (vio_cmo_alloc(viodev, roundup(size, IOMMU_PAGE_SIZE(tbl)))) + goto out_fail; + ret = iommu_map_page(dev, tbl, page, offset, size, dma_get_mask(dev), + direction, attrs); + if (unlikely(ret == DMA_MAPPING_ERROR)) + goto out_deallocate; + return ret; + +out_deallocate: + vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE(tbl))); +out_fail: + atomic_inc(&viodev->cmo.allocs_failed); + return DMA_MAPPING_ERROR; +} + +static void vio_dma_iommu_unmap_page(struct device *dev, dma_addr_t dma_handle, + size_t size, + enum dma_data_direction direction, + unsigned long attrs) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct iommu_table *tbl = get_iommu_table_base(dev); + + iommu_unmap_page(tbl, dma_handle, size, direction, attrs); + vio_cmo_dealloc(viodev, roundup(size, IOMMU_PAGE_SIZE(tbl))); +} + +static int vio_dma_iommu_map_sg(struct device *dev, struct scatterlist *sglist, + int nelems, enum dma_data_direction direction, + unsigned long attrs) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct iommu_table *tbl = get_iommu_table_base(dev); + struct scatterlist *sgl; + int ret, count; + size_t alloc_size = 0; + + for_each_sg(sglist, sgl, nelems, count) + alloc_size += roundup(sgl->length, IOMMU_PAGE_SIZE(tbl)); + + ret = vio_cmo_alloc(viodev, alloc_size); + if (ret) + goto out_fail; + ret = ppc_iommu_map_sg(dev, tbl, sglist, nelems, dma_get_mask(dev), + direction, attrs); + if (unlikely(!ret)) + goto out_deallocate; + + for_each_sg(sglist, sgl, ret, count) + alloc_size -= roundup(sgl->dma_length, IOMMU_PAGE_SIZE(tbl)); + if (alloc_size) + vio_cmo_dealloc(viodev, alloc_size); + return ret; + +out_deallocate: + vio_cmo_dealloc(viodev, alloc_size); +out_fail: + atomic_inc(&viodev->cmo.allocs_failed); + return ret; +} + +static void vio_dma_iommu_unmap_sg(struct device *dev, + struct scatterlist *sglist, int nelems, + enum dma_data_direction direction, + unsigned long attrs) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct iommu_table *tbl = get_iommu_table_base(dev); + struct scatterlist *sgl; + size_t alloc_size = 0; + int count; + + for_each_sg(sglist, sgl, nelems, count) + alloc_size += roundup(sgl->dma_length, IOMMU_PAGE_SIZE(tbl)); + + ppc_iommu_unmap_sg(tbl, sglist, nelems, direction, attrs); + vio_cmo_dealloc(viodev, alloc_size); +} + +static const struct dma_map_ops vio_dma_mapping_ops = { + .alloc = vio_dma_iommu_alloc_coherent, + .free = vio_dma_iommu_free_coherent, + .map_sg = vio_dma_iommu_map_sg, + .unmap_sg = vio_dma_iommu_unmap_sg, + .map_page = vio_dma_iommu_map_page, + .unmap_page = vio_dma_iommu_unmap_page, + .dma_supported = dma_iommu_dma_supported, + .get_required_mask = dma_iommu_get_required_mask, + .mmap = dma_common_mmap, + .get_sgtable = dma_common_get_sgtable, + .alloc_pages = dma_common_alloc_pages, + .free_pages = dma_common_free_pages, +}; + +/** + * vio_cmo_set_dev_desired - Set desired entitlement for a device + * + * @viodev: struct vio_dev for device to alter + * @desired: new desired entitlement level in bytes + * + * For use by devices to request a change to their entitlement at runtime or + * through sysfs. The desired entitlement level is changed and a balancing + * of system resources is scheduled to run in the future. + */ +void vio_cmo_set_dev_desired(struct vio_dev *viodev, size_t desired) +{ + unsigned long flags; + struct vio_cmo_dev_entry *dev_ent; + int found = 0; + + if (!firmware_has_feature(FW_FEATURE_CMO)) + return; + + spin_lock_irqsave(&vio_cmo.lock, flags); + if (desired < VIO_CMO_MIN_ENT) + desired = VIO_CMO_MIN_ENT; + + /* + * Changes will not be made for devices not in the device list. + * If it is not in the device list, then no driver is loaded + * for the device and it can not receive entitlement. + */ + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) + if (viodev == dev_ent->viodev) { + found = 1; + break; + } + if (!found) { + spin_unlock_irqrestore(&vio_cmo.lock, flags); + return; + } + + /* Increase/decrease in desired device entitlement */ + if (desired >= viodev->cmo.desired) { + /* Just bump the bus and device values prior to a balance*/ + vio_cmo.desired += desired - viodev->cmo.desired; + viodev->cmo.desired = desired; + } else { + /* Decrease bus and device values for desired entitlement */ + vio_cmo.desired -= viodev->cmo.desired - desired; + viodev->cmo.desired = desired; + /* + * If less entitlement is desired than current entitlement, move + * any reserve memory in the change region to the excess pool. + */ + if (viodev->cmo.entitled > desired) { + vio_cmo.reserve.size -= viodev->cmo.entitled - desired; + vio_cmo.excess.size += viodev->cmo.entitled - desired; + /* + * If entitlement moving from the reserve pool to the + * excess pool is currently unused, add to the excess + * free counter. + */ + if (viodev->cmo.allocated < viodev->cmo.entitled) + vio_cmo.excess.free += viodev->cmo.entitled - + max(viodev->cmo.allocated, desired); + viodev->cmo.entitled = desired; + } + } + schedule_delayed_work(&vio_cmo.balance_q, 0); + spin_unlock_irqrestore(&vio_cmo.lock, flags); +} + +/** + * vio_cmo_bus_probe - Handle CMO specific bus probe activities + * + * @viodev - Pointer to struct vio_dev for device + * + * Determine the devices IO memory entitlement needs, attempting + * to satisfy the system minimum entitlement at first and scheduling + * a balance operation to take care of the rest at a later time. + * + * Returns: 0 on success, -EINVAL when device doesn't support CMO, and + * -ENOMEM when entitlement is not available for device or + * device entry. + * + */ +static int vio_cmo_bus_probe(struct vio_dev *viodev) +{ + struct vio_cmo_dev_entry *dev_ent; + struct device *dev = &viodev->dev; + struct iommu_table *tbl; + struct vio_driver *viodrv = to_vio_driver(dev->driver); + unsigned long flags; + size_t size; + bool dma_capable = false; + + tbl = get_iommu_table_base(dev); + + /* A device requires entitlement if it has a DMA window property */ + switch (viodev->family) { + case VDEVICE: + if (of_get_property(viodev->dev.of_node, + "ibm,my-dma-window", NULL)) + dma_capable = true; + break; + case PFO: + dma_capable = false; + break; + default: + dev_warn(dev, "unknown device family: %d\n", viodev->family); + BUG(); + break; + } + + /* Configure entitlement for the device. */ + if (dma_capable) { + /* Check that the driver is CMO enabled and get desired DMA */ + if (!viodrv->get_desired_dma) { + dev_err(dev, "%s: device driver does not support CMO\n", + __func__); + return -EINVAL; + } + + viodev->cmo.desired = + IOMMU_PAGE_ALIGN(viodrv->get_desired_dma(viodev), tbl); + if (viodev->cmo.desired < VIO_CMO_MIN_ENT) + viodev->cmo.desired = VIO_CMO_MIN_ENT; + size = VIO_CMO_MIN_ENT; + + dev_ent = kmalloc(sizeof(struct vio_cmo_dev_entry), + GFP_KERNEL); + if (!dev_ent) + return -ENOMEM; + + dev_ent->viodev = viodev; + spin_lock_irqsave(&vio_cmo.lock, flags); + list_add(&dev_ent->list, &vio_cmo.device_list); + } else { + viodev->cmo.desired = 0; + size = 0; + spin_lock_irqsave(&vio_cmo.lock, flags); + } + + /* + * If the needs for vio_cmo.min have not changed since they + * were last set, the number of devices in the OF tree has + * been constant and the IO memory for this is already in + * the reserve pool. + */ + if (vio_cmo.min == ((vio_cmo_num_OF_devs() + 1) * + VIO_CMO_MIN_ENT)) { + /* Updated desired entitlement if device requires it */ + if (size) + vio_cmo.desired += (viodev->cmo.desired - + VIO_CMO_MIN_ENT); + } else { + size_t tmp; + + tmp = vio_cmo.spare + vio_cmo.excess.free; + if (tmp < size) { + dev_err(dev, "%s: insufficient free " + "entitlement to add device. " + "Need %lu, have %lu\n", __func__, + size, (vio_cmo.spare + tmp)); + spin_unlock_irqrestore(&vio_cmo.lock, flags); + return -ENOMEM; + } + + /* Use excess pool first to fulfill request */ + tmp = min(size, vio_cmo.excess.free); + vio_cmo.excess.free -= tmp; + vio_cmo.excess.size -= tmp; + vio_cmo.reserve.size += tmp; + + /* Use spare if excess pool was insufficient */ + vio_cmo.spare -= size - tmp; + + /* Update bus accounting */ + vio_cmo.min += size; + vio_cmo.desired += viodev->cmo.desired; + } + spin_unlock_irqrestore(&vio_cmo.lock, flags); + return 0; +} + +/** + * vio_cmo_bus_remove - Handle CMO specific bus removal activities + * + * @viodev - Pointer to struct vio_dev for device + * + * Remove the device from the cmo device list. The minimum entitlement + * will be reserved for the device as long as it is in the system. The + * rest of the entitlement the device had been allocated will be returned + * to the system. + */ +static void vio_cmo_bus_remove(struct vio_dev *viodev) +{ + struct vio_cmo_dev_entry *dev_ent; + unsigned long flags; + size_t tmp; + + spin_lock_irqsave(&vio_cmo.lock, flags); + if (viodev->cmo.allocated) { + dev_err(&viodev->dev, "%s: device had %lu bytes of IO " + "allocated after remove operation.\n", + __func__, viodev->cmo.allocated); + BUG(); + } + + /* + * Remove the device from the device list being maintained for + * CMO enabled devices. + */ + list_for_each_entry(dev_ent, &vio_cmo.device_list, list) + if (viodev == dev_ent->viodev) { + list_del(&dev_ent->list); + kfree(dev_ent); + break; + } + + /* + * Devices may not require any entitlement and they do not need + * to be processed. Otherwise, return the device's entitlement + * back to the pools. + */ + if (viodev->cmo.entitled) { + /* + * This device has not yet left the OF tree, it's + * minimum entitlement remains in vio_cmo.min and + * vio_cmo.desired + */ + vio_cmo.desired -= (viodev->cmo.desired - VIO_CMO_MIN_ENT); + + /* + * Save min allocation for device in reserve as long + * as it exists in OF tree as determined by later + * balance operation + */ + viodev->cmo.entitled -= VIO_CMO_MIN_ENT; + + /* Replenish spare from freed reserve pool */ + if (viodev->cmo.entitled && (vio_cmo.spare < VIO_CMO_MIN_ENT)) { + tmp = min(viodev->cmo.entitled, (VIO_CMO_MIN_ENT - + vio_cmo.spare)); + vio_cmo.spare += tmp; + viodev->cmo.entitled -= tmp; + } + + /* Remaining reserve goes to excess pool */ + vio_cmo.excess.size += viodev->cmo.entitled; + vio_cmo.excess.free += viodev->cmo.entitled; + vio_cmo.reserve.size -= viodev->cmo.entitled; + + /* + * Until the device is removed it will keep a + * minimum entitlement; this will guarantee that + * a module unload/load will result in a success. + */ + viodev->cmo.entitled = VIO_CMO_MIN_ENT; + viodev->cmo.desired = VIO_CMO_MIN_ENT; + atomic_set(&viodev->cmo.allocs_failed, 0); + } + + spin_unlock_irqrestore(&vio_cmo.lock, flags); +} + +static void vio_cmo_set_dma_ops(struct vio_dev *viodev) +{ + set_dma_ops(&viodev->dev, &vio_dma_mapping_ops); +} + +/** + * vio_cmo_bus_init - CMO entitlement initialization at bus init time + * + * Set up the reserve and excess entitlement pools based on available + * system entitlement and the number of devices in the OF tree that + * require entitlement in the reserve pool. + */ +static void vio_cmo_bus_init(void) +{ + struct hvcall_mpp_data mpp_data; + int err; + + memset(&vio_cmo, 0, sizeof(struct vio_cmo)); + spin_lock_init(&vio_cmo.lock); + INIT_LIST_HEAD(&vio_cmo.device_list); + INIT_DELAYED_WORK(&vio_cmo.balance_q, vio_cmo_balance); + + /* Get current system entitlement */ + err = h_get_mpp(&mpp_data); + + /* + * On failure, continue with entitlement set to 0, will panic() + * later when spare is reserved. + */ + if (err != H_SUCCESS) { + printk(KERN_ERR "%s: unable to determine system IO "\ + "entitlement. (%d)\n", __func__, err); + vio_cmo.entitled = 0; + } else { + vio_cmo.entitled = mpp_data.entitled_mem; + } + + /* Set reservation and check against entitlement */ + vio_cmo.spare = VIO_CMO_MIN_ENT; + vio_cmo.reserve.size = vio_cmo.spare; + vio_cmo.reserve.size += (vio_cmo_num_OF_devs() * + VIO_CMO_MIN_ENT); + if (vio_cmo.reserve.size > vio_cmo.entitled) { + printk(KERN_ERR "%s: insufficient system entitlement\n", + __func__); + panic("%s: Insufficient system entitlement", __func__); + } + + /* Set the remaining accounting variables */ + vio_cmo.excess.size = vio_cmo.entitled - vio_cmo.reserve.size; + vio_cmo.excess.free = vio_cmo.excess.size; + vio_cmo.min = vio_cmo.reserve.size; + vio_cmo.desired = vio_cmo.reserve.size; +} + +/* sysfs device functions and data structures for CMO */ + +#define viodev_cmo_rd_attr(name) \ +static ssize_t cmo_##name##_show(struct device *dev, \ + struct device_attribute *attr, \ + char *buf) \ +{ \ + return sprintf(buf, "%lu\n", to_vio_dev(dev)->cmo.name); \ +} + +static ssize_t cmo_allocs_failed_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct vio_dev *viodev = to_vio_dev(dev); + return sprintf(buf, "%d\n", atomic_read(&viodev->cmo.allocs_failed)); +} + +static ssize_t cmo_allocs_failed_store(struct device *dev, + struct device_attribute *attr, const char *buf, size_t count) +{ + struct vio_dev *viodev = to_vio_dev(dev); + atomic_set(&viodev->cmo.allocs_failed, 0); + return count; +} + +static ssize_t cmo_desired_store(struct device *dev, + struct device_attribute *attr, const char *buf, size_t count) +{ + struct vio_dev *viodev = to_vio_dev(dev); + size_t new_desired; + int ret; + + ret = kstrtoul(buf, 10, &new_desired); + if (ret) + return ret; + + vio_cmo_set_dev_desired(viodev, new_desired); + return count; +} + +viodev_cmo_rd_attr(desired); +viodev_cmo_rd_attr(entitled); +viodev_cmo_rd_attr(allocated); + +static ssize_t name_show(struct device *, struct device_attribute *, char *); +static ssize_t devspec_show(struct device *, struct device_attribute *, char *); +static ssize_t modalias_show(struct device *dev, struct device_attribute *attr, + char *buf); + +static struct device_attribute dev_attr_name; +static struct device_attribute dev_attr_devspec; +static struct device_attribute dev_attr_modalias; + +static DEVICE_ATTR_RO(cmo_entitled); +static DEVICE_ATTR_RO(cmo_allocated); +static DEVICE_ATTR_RW(cmo_desired); +static DEVICE_ATTR_RW(cmo_allocs_failed); + +static struct attribute *vio_cmo_dev_attrs[] = { + &dev_attr_name.attr, + &dev_attr_devspec.attr, + &dev_attr_modalias.attr, + &dev_attr_cmo_entitled.attr, + &dev_attr_cmo_allocated.attr, + &dev_attr_cmo_desired.attr, + &dev_attr_cmo_allocs_failed.attr, + NULL, +}; +ATTRIBUTE_GROUPS(vio_cmo_dev); + +/* sysfs bus functions and data structures for CMO */ + +#define viobus_cmo_rd_attr(name) \ +static ssize_t cmo_bus_##name##_show(const struct bus_type *bt, char *buf) \ +{ \ + return sprintf(buf, "%lu\n", vio_cmo.name); \ +} \ +static struct bus_attribute bus_attr_cmo_bus_##name = \ + __ATTR(cmo_##name, S_IRUGO, cmo_bus_##name##_show, NULL) + +#define viobus_cmo_pool_rd_attr(name, var) \ +static ssize_t \ +cmo_##name##_##var##_show(const struct bus_type *bt, char *buf) \ +{ \ + return sprintf(buf, "%lu\n", vio_cmo.name.var); \ +} \ +static BUS_ATTR_RO(cmo_##name##_##var) + +viobus_cmo_rd_attr(entitled); +viobus_cmo_rd_attr(spare); +viobus_cmo_rd_attr(min); +viobus_cmo_rd_attr(desired); +viobus_cmo_rd_attr(curr); +viobus_cmo_pool_rd_attr(reserve, size); +viobus_cmo_pool_rd_attr(excess, size); +viobus_cmo_pool_rd_attr(excess, free); + +static ssize_t cmo_high_show(const struct bus_type *bt, char *buf) +{ + return sprintf(buf, "%lu\n", vio_cmo.high); +} + +static ssize_t cmo_high_store(const struct bus_type *bt, const char *buf, + size_t count) +{ + unsigned long flags; + + spin_lock_irqsave(&vio_cmo.lock, flags); + vio_cmo.high = vio_cmo.curr; + spin_unlock_irqrestore(&vio_cmo.lock, flags); + + return count; +} +static BUS_ATTR_RW(cmo_high); + +static struct attribute *vio_bus_attrs[] = { + &bus_attr_cmo_bus_entitled.attr, + &bus_attr_cmo_bus_spare.attr, + &bus_attr_cmo_bus_min.attr, + &bus_attr_cmo_bus_desired.attr, + &bus_attr_cmo_bus_curr.attr, + &bus_attr_cmo_high.attr, + &bus_attr_cmo_reserve_size.attr, + &bus_attr_cmo_excess_size.attr, + &bus_attr_cmo_excess_free.attr, + NULL, +}; +ATTRIBUTE_GROUPS(vio_bus); + +static void __init vio_cmo_sysfs_init(void) +{ + vio_bus_type.dev_groups = vio_cmo_dev_groups; + vio_bus_type.bus_groups = vio_bus_groups; +} +#else /* CONFIG_PPC_SMLPAR */ +int vio_cmo_entitlement_update(size_t new_entitlement) { return 0; } +void vio_cmo_set_dev_desired(struct vio_dev *viodev, size_t desired) {} +static int vio_cmo_bus_probe(struct vio_dev *viodev) { return 0; } +static void vio_cmo_bus_remove(struct vio_dev *viodev) {} +static void vio_cmo_set_dma_ops(struct vio_dev *viodev) {} +static void vio_cmo_bus_init(void) {} +static void __init vio_cmo_sysfs_init(void) { } +#endif /* CONFIG_PPC_SMLPAR */ +EXPORT_SYMBOL(vio_cmo_entitlement_update); +EXPORT_SYMBOL(vio_cmo_set_dev_desired); + + +/* + * Platform Facilities Option (PFO) support + */ + +/** + * vio_h_cop_sync - Perform a synchronous PFO co-processor operation + * + * @vdev - Pointer to a struct vio_dev for device + * @op - Pointer to a struct vio_pfo_op for the operation parameters + * + * Calls the hypervisor to synchronously perform the PFO operation + * described in @op. In the case of a busy response from the hypervisor, + * the operation will be re-submitted indefinitely unless a non-zero timeout + * is specified or an error occurs. The timeout places a limit on when to + * stop re-submitting a operation, the total time can be exceeded if an + * operation is in progress. + * + * If op->hcall_ret is not NULL, this will be set to the return from the + * last h_cop_op call or it will be 0 if an error not involving the h_call + * was encountered. + * + * Returns: + * 0 on success, + * -EINVAL if the h_call fails due to an invalid parameter, + * -E2BIG if the h_call can not be performed synchronously, + * -EBUSY if a timeout is specified and has elapsed, + * -EACCES if the memory area for data/status has been rescinded, or + * -EPERM if a hardware fault has been indicated + */ +int vio_h_cop_sync(struct vio_dev *vdev, struct vio_pfo_op *op) +{ + struct device *dev = &vdev->dev; + unsigned long deadline = 0; + long hret = 0; + int ret = 0; + + if (op->timeout) + deadline = jiffies + msecs_to_jiffies(op->timeout); + + while (true) { + hret = plpar_hcall_norets(H_COP, op->flags, + vdev->resource_id, + op->in, op->inlen, op->out, + op->outlen, op->csbcpb); + + if (hret == H_SUCCESS || + (hret != H_NOT_ENOUGH_RESOURCES && + hret != H_BUSY && hret != H_RESOURCE) || + (op->timeout && time_after(deadline, jiffies))) + break; + + dev_dbg(dev, "%s: hcall ret(%ld), retrying.\n", __func__, hret); + } + + switch (hret) { + case H_SUCCESS: + ret = 0; + break; + case H_OP_MODE: + case H_TOO_BIG: + ret = -E2BIG; + break; + case H_RESCINDED: + ret = -EACCES; + break; + case H_HARDWARE: + ret = -EPERM; + break; + case H_NOT_ENOUGH_RESOURCES: + case H_RESOURCE: + case H_BUSY: + ret = -EBUSY; + break; + default: + ret = -EINVAL; + break; + } + + if (ret) + dev_dbg(dev, "%s: Sync h_cop_op failure (ret:%d) (hret:%ld)\n", + __func__, ret, hret); + + op->hcall_err = hret; + return ret; +} +EXPORT_SYMBOL(vio_h_cop_sync); + +static struct iommu_table *vio_build_iommu_table(struct vio_dev *dev) +{ + const __be32 *dma_window; + struct iommu_table *tbl; + unsigned long offset, size; + + dma_window = of_get_property(dev->dev.of_node, + "ibm,my-dma-window", NULL); + if (!dma_window) + return NULL; + + tbl = kzalloc(sizeof(*tbl), GFP_KERNEL); + if (tbl == NULL) + return NULL; + + kref_init(&tbl->it_kref); + + of_parse_dma_window(dev->dev.of_node, dma_window, + &tbl->it_index, &offset, &size); + + /* TCE table size - measured in tce entries */ + tbl->it_page_shift = IOMMU_PAGE_SHIFT_4K; + tbl->it_size = size >> tbl->it_page_shift; + /* offset for VIO should always be 0 */ + tbl->it_offset = offset >> tbl->it_page_shift; + tbl->it_busno = 0; + tbl->it_type = TCE_VB; + tbl->it_blocksize = 16; + + if (firmware_has_feature(FW_FEATURE_LPAR)) + tbl->it_ops = &iommu_table_lpar_multi_ops; + else + tbl->it_ops = &iommu_table_pseries_ops; + + return iommu_init_table(tbl, -1, 0, 0); +} + +/** + * vio_match_device: - Tell if a VIO device has a matching + * VIO device id structure. + * @ids: array of VIO device id structures to search in + * @dev: the VIO device structure to match against + * + * Used by a driver to check whether a VIO device present in the + * system is in its list of supported devices. Returns the matching + * vio_device_id structure or NULL if there is no match. + */ +static const struct vio_device_id *vio_match_device( + const struct vio_device_id *ids, const struct vio_dev *dev) +{ + while (ids->type[0] != '\0') { + if ((strncmp(dev->type, ids->type, strlen(ids->type)) == 0) && + of_device_is_compatible(dev->dev.of_node, + ids->compat)) + return ids; + ids++; + } + return NULL; +} + +/* + * Convert from struct device to struct vio_dev and pass to driver. + * dev->driver has already been set by generic code because vio_bus_match + * succeeded. + */ +static int vio_bus_probe(struct device *dev) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct vio_driver *viodrv = to_vio_driver(dev->driver); + const struct vio_device_id *id; + int error = -ENODEV; + + if (!viodrv->probe) + return error; + + id = vio_match_device(viodrv->id_table, viodev); + if (id) { + memset(&viodev->cmo, 0, sizeof(viodev->cmo)); + if (firmware_has_feature(FW_FEATURE_CMO)) { + error = vio_cmo_bus_probe(viodev); + if (error) + return error; + } + error = viodrv->probe(viodev, id); + if (error && firmware_has_feature(FW_FEATURE_CMO)) + vio_cmo_bus_remove(viodev); + } + + return error; +} + +/* convert from struct device to struct vio_dev and pass to driver. */ +static void vio_bus_remove(struct device *dev) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct vio_driver *viodrv = to_vio_driver(dev->driver); + struct device *devptr; + + /* + * Hold a reference to the device after the remove function is called + * to allow for CMO accounting cleanup for the device. + */ + devptr = get_device(dev); + + if (viodrv->remove) + viodrv->remove(viodev); + + if (firmware_has_feature(FW_FEATURE_CMO)) + vio_cmo_bus_remove(viodev); + + put_device(devptr); +} + +static void vio_bus_shutdown(struct device *dev) +{ + struct vio_dev *viodev = to_vio_dev(dev); + struct vio_driver *viodrv; + + if (dev->driver) { + viodrv = to_vio_driver(dev->driver); + if (viodrv->shutdown) + viodrv->shutdown(viodev); + else if (kexec_in_progress) + vio_bus_remove(dev); + } +} + +/** + * vio_register_driver: - Register a new vio driver + * @viodrv: The vio_driver structure to be registered. + */ +int __vio_register_driver(struct vio_driver *viodrv, struct module *owner, + const char *mod_name) +{ + // vio_bus_type is only initialised for pseries + if (!machine_is(pseries)) + return -ENODEV; + + pr_debug("%s: driver %s registering\n", __func__, viodrv->name); + + /* fill in 'struct driver' fields */ + viodrv->driver.name = viodrv->name; + viodrv->driver.pm = viodrv->pm; + viodrv->driver.bus = &vio_bus_type; + viodrv->driver.owner = owner; + viodrv->driver.mod_name = mod_name; + + return driver_register(&viodrv->driver); +} +EXPORT_SYMBOL(__vio_register_driver); + +/** + * vio_unregister_driver - Remove registration of vio driver. + * @viodrv: The vio_driver struct to be removed form registration + */ +void vio_unregister_driver(struct vio_driver *viodrv) +{ + driver_unregister(&viodrv->driver); +} +EXPORT_SYMBOL(vio_unregister_driver); + +/* vio_dev refcount hit 0 */ +static void vio_dev_release(struct device *dev) +{ + struct iommu_table *tbl = get_iommu_table_base(dev); + + if (tbl) + iommu_tce_table_put(tbl); + of_node_put(dev->of_node); + kfree(to_vio_dev(dev)); +} + +/** + * vio_register_device_node: - Register a new vio device. + * @of_node: The OF node for this device. + * + * Creates and initializes a vio_dev structure from the data in + * of_node and adds it to the list of virtual devices. + * Returns a pointer to the created vio_dev or NULL if node has + * NULL device_type or compatible fields. + */ +struct vio_dev *vio_register_device_node(struct device_node *of_node) +{ + struct vio_dev *viodev; + struct device_node *parent_node; + const __be32 *prop; + enum vio_dev_family family; + + /* + * Determine if this node is a under the /vdevice node or under the + * /ibm,platform-facilities node. This decides the device's family. + */ + parent_node = of_get_parent(of_node); + if (parent_node) { + if (of_node_is_type(parent_node, "ibm,platform-facilities")) + family = PFO; + else if (of_node_is_type(parent_node, "vdevice")) + family = VDEVICE; + else { + pr_warn("%s: parent(%pOF) of %pOFn not recognized.\n", + __func__, + parent_node, + of_node); + of_node_put(parent_node); + return NULL; + } + of_node_put(parent_node); + } else { + pr_warn("%s: could not determine the parent of node %pOFn.\n", + __func__, of_node); + return NULL; + } + + if (family == PFO) { + if (of_property_read_bool(of_node, "interrupt-controller")) { + pr_debug("%s: Skipping the interrupt controller %pOFn.\n", + __func__, of_node); + return NULL; + } + } + + /* allocate a vio_dev for this node */ + viodev = kzalloc(sizeof(struct vio_dev), GFP_KERNEL); + if (viodev == NULL) { + pr_warn("%s: allocation failure for VIO device.\n", __func__); + return NULL; + } + + /* we need the 'device_type' property, in order to match with drivers */ + viodev->family = family; + if (viodev->family == VDEVICE) { + unsigned int unit_address; + + viodev->type = of_node_get_device_type(of_node); + if (!viodev->type) { + pr_warn("%s: node %pOFn is missing the 'device_type' " + "property.\n", __func__, of_node); + goto out; + } + + prop = of_get_property(of_node, "reg", NULL); + if (prop == NULL) { + pr_warn("%s: node %pOFn missing 'reg'\n", + __func__, of_node); + goto out; + } + unit_address = of_read_number(prop, 1); + dev_set_name(&viodev->dev, "%x", unit_address); + viodev->irq = irq_of_parse_and_map(of_node, 0); + viodev->unit_address = unit_address; + } else { + /* PFO devices need their resource_id for submitting COP_OPs + * This is an optional field for devices, but is required when + * performing synchronous ops */ + prop = of_get_property(of_node, "ibm,resource-id", NULL); + if (prop != NULL) + viodev->resource_id = of_read_number(prop, 1); + + dev_set_name(&viodev->dev, "%pOFn", of_node); + viodev->type = dev_name(&viodev->dev); + viodev->irq = 0; + } + + viodev->name = of_node->name; + viodev->dev.of_node = of_node_get(of_node); + + set_dev_node(&viodev->dev, of_node_to_nid(of_node)); + + /* init generic 'struct device' fields: */ + viodev->dev.parent = &vio_bus_device.dev; + viodev->dev.bus = &vio_bus_type; + viodev->dev.release = vio_dev_release; + + if (of_property_present(viodev->dev.of_node, "ibm,my-dma-window")) { + if (firmware_has_feature(FW_FEATURE_CMO)) + vio_cmo_set_dma_ops(viodev); + else + set_dma_ops(&viodev->dev, &dma_iommu_ops); + + set_iommu_table_base(&viodev->dev, + vio_build_iommu_table(viodev)); + + /* needed to ensure proper operation of coherent allocations + * later, in case driver doesn't set it explicitly */ + viodev->dev.coherent_dma_mask = DMA_BIT_MASK(64); + viodev->dev.dma_mask = &viodev->dev.coherent_dma_mask; + } + + /* register with generic device framework */ + if (device_register(&viodev->dev)) { + printk(KERN_ERR "%s: failed to register device %s\n", + __func__, dev_name(&viodev->dev)); + put_device(&viodev->dev); + return NULL; + } + + return viodev; + +out: /* Use this exit point for any return prior to device_register */ + kfree(viodev); + + return NULL; +} +EXPORT_SYMBOL(vio_register_device_node); + +/* + * vio_bus_scan_for_devices - Scan OF and register each child device + * @root_name - OF node name for the root of the subtree to search. + * This must be non-NULL + * + * Starting from the root node provide, register the device node for + * each child beneath the root. + */ +static void __init vio_bus_scan_register_devices(char *root_name) +{ + struct device_node *node_root, *node_child; + + if (!root_name) + return; + + node_root = of_find_node_by_name(NULL, root_name); + if (node_root) { + + /* + * Create struct vio_devices for each virtual device in + * the device tree. Drivers will associate with them later. + */ + node_child = of_get_next_child(node_root, NULL); + while (node_child) { + vio_register_device_node(node_child); + node_child = of_get_next_child(node_root, node_child); + } + of_node_put(node_root); + } +} + +/** + * vio_bus_init: - Initialize the virtual IO bus + */ +static int __init vio_bus_init(void) +{ + int err; + + if (firmware_has_feature(FW_FEATURE_CMO)) + vio_cmo_sysfs_init(); + + err = bus_register(&vio_bus_type); + if (err) { + printk(KERN_ERR "failed to register VIO bus\n"); + return err; + } + + /* + * The fake parent of all vio devices, just to give us + * a nice directory + */ + err = device_register(&vio_bus_device.dev); + if (err) { + printk(KERN_WARNING "%s: device_register returned %i\n", + __func__, err); + return err; + } + + if (firmware_has_feature(FW_FEATURE_CMO)) + vio_cmo_bus_init(); + + return 0; +} +machine_postcore_initcall(pseries, vio_bus_init); + +static int __init vio_device_init(void) +{ + vio_bus_scan_register_devices("vdevice"); + vio_bus_scan_register_devices("ibm,platform-facilities"); + + return 0; +} +machine_device_initcall(pseries, vio_device_init); + +static ssize_t name_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "%s\n", to_vio_dev(dev)->name); +} +static DEVICE_ATTR_RO(name); + +static ssize_t devspec_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct device_node *of_node = dev->of_node; + + return sprintf(buf, "%pOF\n", of_node); +} +static DEVICE_ATTR_RO(devspec); + +static ssize_t modalias_show(struct device *dev, struct device_attribute *attr, + char *buf) +{ + const struct vio_dev *vio_dev = to_vio_dev(dev); + struct device_node *dn; + const char *cp; + + dn = dev->of_node; + if (!dn) { + strcpy(buf, "\n"); + return strlen(buf); + } + cp = of_get_property(dn, "compatible", NULL); + if (!cp) { + strcpy(buf, "\n"); + return strlen(buf); + } + + return sprintf(buf, "vio:T%sS%s\n", vio_dev->type, cp); +} +static DEVICE_ATTR_RO(modalias); + +static struct attribute *vio_dev_attrs[] = { + &dev_attr_name.attr, + &dev_attr_devspec.attr, + &dev_attr_modalias.attr, + NULL, +}; +ATTRIBUTE_GROUPS(vio_dev); + +void vio_unregister_device(struct vio_dev *viodev) +{ + device_unregister(&viodev->dev); + if (viodev->family == VDEVICE) + irq_dispose_mapping(viodev->irq); +} +EXPORT_SYMBOL(vio_unregister_device); + +static int vio_bus_match(struct device *dev, struct device_driver *drv) +{ + const struct vio_dev *vio_dev = to_vio_dev(dev); + struct vio_driver *vio_drv = to_vio_driver(drv); + const struct vio_device_id *ids = vio_drv->id_table; + + return (ids != NULL) && (vio_match_device(ids, vio_dev) != NULL); +} + +static int vio_hotplug(const struct device *dev, struct kobj_uevent_env *env) +{ + const struct vio_dev *vio_dev = to_vio_dev(dev); + const struct device_node *dn; + const char *cp; + + dn = dev->of_node; + if (!dn) + return -ENODEV; + cp = of_get_property(dn, "compatible", NULL); + if (!cp) + return -ENODEV; + + add_uevent_var(env, "MODALIAS=vio:T%sS%s", vio_dev->type, cp); + return 0; +} + +struct bus_type vio_bus_type = { + .name = "vio", + .dev_groups = vio_dev_groups, + .uevent = vio_hotplug, + .match = vio_bus_match, + .probe = vio_bus_probe, + .remove = vio_bus_remove, + .shutdown = vio_bus_shutdown, +}; + +/** + * vio_get_attribute: - get attribute for virtual device + * @vdev: The vio device to get property. + * @which: The property/attribute to be extracted. + * @length: Pointer to length of returned data size (unused if NULL). + * + * Calls prom.c's of_get_property() to return the value of the + * attribute specified by @which +*/ +const void *vio_get_attribute(struct vio_dev *vdev, char *which, int *length) +{ + return of_get_property(vdev->dev.of_node, which, length); +} +EXPORT_SYMBOL(vio_get_attribute); + +/* vio_find_name() - internal because only vio.c knows how we formatted the + * kobject name + */ +static struct vio_dev *vio_find_name(const char *name) +{ + struct device *found; + + found = bus_find_device_by_name(&vio_bus_type, NULL, name); + if (!found) + return NULL; + + return to_vio_dev(found); +} + +/** + * vio_find_node - find an already-registered vio_dev + * @vnode: device_node of the virtual device we're looking for + * + * Takes a reference to the embedded struct device which needs to be dropped + * after use. + */ +struct vio_dev *vio_find_node(struct device_node *vnode) +{ + char kobj_name[20]; + struct device_node *vnode_parent; + + vnode_parent = of_get_parent(vnode); + if (!vnode_parent) + return NULL; + + /* construct the kobject name from the device node */ + if (of_node_is_type(vnode_parent, "vdevice")) { + const __be32 *prop; + + prop = of_get_property(vnode, "reg", NULL); + if (!prop) + goto out; + snprintf(kobj_name, sizeof(kobj_name), "%x", + (uint32_t)of_read_number(prop, 1)); + } else if (of_node_is_type(vnode_parent, "ibm,platform-facilities")) + snprintf(kobj_name, sizeof(kobj_name), "%pOFn", vnode); + else + goto out; + + of_node_put(vnode_parent); + return vio_find_name(kobj_name); +out: + of_node_put(vnode_parent); + return NULL; +} +EXPORT_SYMBOL(vio_find_node); + +int vio_enable_interrupts(struct vio_dev *dev) +{ + int rc = h_vio_signal(dev->unit_address, VIO_IRQ_ENABLE); + if (rc != H_SUCCESS) + printk(KERN_ERR "vio: Error 0x%x enabling interrupts\n", rc); + return rc; +} +EXPORT_SYMBOL(vio_enable_interrupts); + +int vio_disable_interrupts(struct vio_dev *dev) +{ + int rc = h_vio_signal(dev->unit_address, VIO_IRQ_DISABLE); + if (rc != H_SUCCESS) + printk(KERN_ERR "vio: Error 0x%x disabling interrupts\n", rc); + return rc; +} +EXPORT_SYMBOL(vio_disable_interrupts); + +static int __init vio_init(void) +{ + dma_debug_add_bus(&vio_bus_type); + return 0; +} +machine_fs_initcall(pseries, vio_init); diff --git a/arch/powerpc/platforms/pseries/vphn.c b/arch/powerpc/platforms/pseries/vphn.c new file mode 100644 index 000000000..3f85ece3c --- /dev/null +++ b/arch/powerpc/platforms/pseries/vphn.c @@ -0,0 +1,90 @@ +// SPDX-License-Identifier: GPL-2.0 +#include <asm/byteorder.h> +#include <asm/vphn.h> + +/* + * The associativity domain numbers are returned from the hypervisor as a + * stream of mixed 16-bit and 32-bit fields. The stream is terminated by the + * special value of "all ones" (aka. 0xffff) and its size may not exceed 48 + * bytes. + * + * --- 16-bit fields --> + * _________________________ + * | 0 | 1 | 2 | 3 | be_packed[0] + * ------+-----+-----+------ + * _________________________ + * | 4 | 5 | 6 | 7 | be_packed[1] + * ------------------------- + * ... + * _________________________ + * | 20 | 21 | 22 | 23 | be_packed[5] + * ------------------------- + * + * Convert to the sequence they would appear in the ibm,associativity property. + */ +static int vphn_unpack_associativity(const long *packed, __be32 *unpacked) +{ + __be64 be_packed[VPHN_REGISTER_COUNT]; + int i, nr_assoc_doms = 0; + const __be16 *field = (const __be16 *) be_packed; + u16 last = 0; + bool is_32bit = false; + +#define VPHN_FIELD_UNUSED (0xffff) +#define VPHN_FIELD_MSB (0x8000) +#define VPHN_FIELD_MASK (~VPHN_FIELD_MSB) + + /* Let's fix the values returned by plpar_hcall9() */ + for (i = 0; i < VPHN_REGISTER_COUNT; i++) + be_packed[i] = cpu_to_be64(packed[i]); + + for (i = 1; i < VPHN_ASSOC_BUFSIZE; i++) { + u16 new = be16_to_cpup(field++); + + if (is_32bit) { + /* + * Let's concatenate the 16 bits of this field to the + * 15 lower bits of the previous field + */ + unpacked[++nr_assoc_doms] = + cpu_to_be32(last << 16 | new); + is_32bit = false; + } else if (new == VPHN_FIELD_UNUSED) + /* This is the list terminator */ + break; + else if (new & VPHN_FIELD_MSB) { + /* Data is in the lower 15 bits of this field */ + unpacked[++nr_assoc_doms] = + cpu_to_be32(new & VPHN_FIELD_MASK); + } else { + /* + * Data is in the lower 15 bits of this field + * concatenated with the next 16 bit field + */ + last = new; + is_32bit = true; + } + } + + /* The first cell contains the length of the property */ + unpacked[0] = cpu_to_be32(nr_assoc_doms); + + return nr_assoc_doms; +} + +/* NOTE: This file is included by a selftest and built in userspace. */ +#ifdef __KERNEL__ +#include <asm/hvcall.h> + +long hcall_vphn(unsigned long cpu, u64 flags, __be32 *associativity) +{ + long rc; + long retbuf[PLPAR_HCALL9_BUFSIZE] = {0}; + + rc = plpar_hcall9(H_HOME_NODE_ASSOCIATIVITY, retbuf, flags, cpu); + if (rc == H_SUCCESS) + vphn_unpack_associativity(retbuf, associativity); + + return rc; +} +#endif |